From 130f43e4b87d17ba9d1c68234e26d1180f4bb9a1 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Tue, 14 May 2024 19:15:35 +0300 Subject: [PATCH 01/17] scripts : sync ggml-rpc --- scripts/sync-ggml-am.sh | 4 ++++ scripts/sync-ggml.sh | 2 ++ 2 files changed, 6 insertions(+) diff --git a/scripts/sync-ggml-am.sh b/scripts/sync-ggml-am.sh index 70ff16d0df2..54c243a15db 100755 --- a/scripts/sync-ggml-am.sh +++ b/scripts/sync-ggml-am.sh @@ -117,6 +117,8 @@ if [ -f $SRC_WHISPER/ggml-src.patch ]; then # src/ggml-opencl.h -> ggml-opencl.h # src/ggml-quants.c -> ggml-quants.c # src/ggml-quants.h -> ggml-quants.h + # src/ggml-rpc.cpp -> ggml-rpc.cpp + # src/ggml-rpc.h -> ggml-rpc.h # src/ggml-sycl.cpp -> ggml-sycl.cpp # src/ggml-sycl.h -> ggml-sycl.h # src/ggml-vulkan.cpp -> ggml-vulkan.cpp @@ -160,6 +162,8 @@ if [ -f $SRC_WHISPER/ggml-src.patch ]; then -e 's/src\/ggml-opencl\.h/ggml-opencl.h/g' \ -e 's/src\/ggml-quants\.c/ggml-quants.c/g' \ -e 's/src\/ggml-quants\.h/ggml-quants.h/g' \ + -e 's/src\/ggml-rpc\.cpp/ggml-rpc.cpp/g' \ + -e 's/src\/ggml-rpc\.h/ggml-rpc.h/g' \ -e 's/src\/ggml-sycl\.cpp/ggml-sycl.cpp/g' \ -e 's/src\/ggml-sycl\.h/ggml-sycl.h/g' \ -e 's/src\/ggml-vulkan\.cpp/ggml-vulkan.cpp/g' \ diff --git a/scripts/sync-ggml.sh b/scripts/sync-ggml.sh index 2efffcd213c..1b0f2045cf3 100755 --- a/scripts/sync-ggml.sh +++ b/scripts/sync-ggml.sh @@ -20,6 +20,8 @@ cp -rpv ../ggml/src/ggml-opencl.cpp ./ggml-opencl.cpp cp -rpv ../ggml/src/ggml-opencl.h ./ggml-opencl.h cp -rpv ../ggml/src/ggml-quants.c ./ggml-quants.c cp -rpv ../ggml/src/ggml-quants.h ./ggml-quants.h +cp -rpv ../ggml/src/ggml-rpc.cpp ./ggml-rpc.cpp +cp -rpv ../ggml/src/ggml-rpc.h ./ggml-rpc.h cp -rpv ../ggml/src/ggml-sycl.cpp ./ggml-sycl.cpp cp -rpv ../ggml/src/ggml-sycl.h ./ggml-sycl.h cp -rpv ../ggml/src/ggml-vulkan.cpp ./ggml-vulkan.cpp From e57e95eb0d3bdba42bbf057c888f6ff819a5f59b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Johannes=20G=C3=A4=C3=9Fler?= Date: Sun, 12 May 2024 19:40:45 +0200 Subject: [PATCH 02/17] CUDA: add FP32 FlashAttention vector kernel (llama/7188) * CUDA: add FP32 FlashAttention vector kernel * fixup! CUDA: add FP32 FlashAttention vector kernel * fixup! fixup! CUDA: add FP32 FlashAttention vector kernel * fixup! fixup! fixup! CUDA: add FP32 FlashAttention vector kernel --- ggml-cuda.cu | 11 +- ggml-cuda/common.cuh | 4 + ggml-cuda/fattn-common.cuh | 47 ++++ ggml-cuda/fattn-vec-f16.cu | 430 +++++++++++++++++++++++++++++++++ ggml-cuda/fattn-vec-f16.cuh | 5 + ggml-cuda/fattn-vec-f32.cu | 384 +++++++++++++++++++++++++++++ ggml-cuda/fattn-vec-f32.cuh | 3 + ggml-cuda/fattn.cu | 468 ++---------------------------------- 8 files changed, 898 insertions(+), 454 deletions(-) create mode 100644 ggml-cuda/fattn-common.cuh create mode 100644 ggml-cuda/fattn-vec-f16.cu create mode 100644 ggml-cuda/fattn-vec-f16.cuh create mode 100644 ggml-cuda/fattn-vec-f32.cu create mode 100644 ggml-cuda/fattn-vec-f32.cuh diff --git a/ggml-cuda.cu b/ggml-cuda.cu index 5b6c9091924..75a2ad48087 100644 --- a/ggml-cuda.cu +++ b/ggml-cuda.cu @@ -2713,6 +2713,7 @@ GGML_CALL static enum ggml_status ggml_backend_cuda_graph_compute(ggml_backend_t } GGML_CALL static bool ggml_backend_cuda_supports_op(ggml_backend_t backend, const ggml_tensor * op) { + ggml_backend_cuda_context * cuda_ctx = (ggml_backend_cuda_context *) backend->context; switch (op->op) { case GGML_OP_UNARY: switch (ggml_get_unary_op(op)) { @@ -2840,8 +2841,16 @@ GGML_CALL static bool ggml_backend_cuda_supports_op(ggml_backend_t backend, cons case GGML_OP_ARANGE: case GGML_OP_TIMESTEP_EMBEDDING: case GGML_OP_LEAKY_RELU: - case GGML_OP_FLASH_ATTN_EXT: return true; + case GGML_OP_FLASH_ATTN_EXT: +#if defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) + return op->src[0]->ne[0] == 64 || op->src[0]->ne[0] == 128; +#else + if (op->src[0]->ne[0] == 64 || op->src[0]->ne[0] == 128) { + return true; + } + return ggml_cuda_info().devices[cuda_ctx->device].cc >= CC_VOLTA; +#endif // defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__) default: return false; } diff --git a/ggml-cuda/common.cuh b/ggml-cuda/common.cuh index 44e67e040e1..b6f0bc36a4f 100644 --- a/ggml-cuda/common.cuh +++ b/ggml-cuda/common.cuh @@ -321,6 +321,10 @@ static __device__ __forceinline__ int __dp4a(const int a, const int b, int c) { #define FP16_MMA_AVAILABLE !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) && __CUDA_ARCH__ >= CC_VOLTA +static bool fast_fp16_available(const int cc) { + return cc >= CC_PASCAL && cc != 610; +} + static bool fp16_mma_available(const int cc) { return cc < CC_OFFSET_AMD && cc >= CC_VOLTA; } diff --git a/ggml-cuda/fattn-common.cuh b/ggml-cuda/fattn-common.cuh new file mode 100644 index 00000000000..33f640691ad --- /dev/null +++ b/ggml-cuda/fattn-common.cuh @@ -0,0 +1,47 @@ +#define FATTN_KQ_STRIDE 256 +#define HALF_MAX_HALF __float2half(65504.0f/2) // Use neg. of this instead of -INFINITY to initialize KQ max vals to avoid NaN upon subtraction. +#define SOFTMAX_FTZ_THRESHOLD -20.0f // Softmax exp. of values smaller than this are flushed to zero to avoid NaNs. + +template // D == head size +#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +__launch_bounds__(D, 1) +#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +static __global__ void flash_attn_combine_results( + const float * __restrict__ VKQ_parts, + const float2 * __restrict__ VKQ_meta, + float * __restrict__ dst) { + VKQ_parts += parallel_blocks*D * gridDim.y*blockIdx.x; + VKQ_meta += parallel_blocks * gridDim.y*blockIdx.x; + dst += D * gridDim.y*blockIdx.x; + + const int tid = threadIdx.x; + __builtin_assume(tid < D); + + __shared__ float2 meta[parallel_blocks]; + if (tid < 2*parallel_blocks) { + ((float *) meta)[threadIdx.x] = ((const float *)VKQ_meta) [blockIdx.y*(2*parallel_blocks) + tid]; + } + + __syncthreads(); + + float kqmax = meta[0].x; +#pragma unroll + for (int l = 1; l < parallel_blocks; ++l) { + kqmax = max(kqmax, meta[l].x); + } + + float VKQ_numerator = 0.0f; + float VKQ_denominator = 0.0f; +#pragma unroll + for (int l = 0; l < parallel_blocks; ++l) { + const float diff = meta[l].x - kqmax; + const float KQ_max_scale = expf(diff); + const uint32_t ftz_mask = 0xFFFFFFFF * (diff > SOFTMAX_FTZ_THRESHOLD); + *((uint32_t *) &KQ_max_scale) &= ftz_mask; + + VKQ_numerator += KQ_max_scale * VKQ_parts[l*gridDim.y*D + blockIdx.y*D + tid]; + VKQ_denominator += KQ_max_scale * meta[l].y; + } + + dst[blockIdx.y*D + tid] = VKQ_numerator / VKQ_denominator; +} diff --git a/ggml-cuda/fattn-vec-f16.cu b/ggml-cuda/fattn-vec-f16.cu new file mode 100644 index 00000000000..cbf5f7835f8 --- /dev/null +++ b/ggml-cuda/fattn-vec-f16.cu @@ -0,0 +1,430 @@ +#include "common.cuh" +#include "fattn-common.cuh" +#include "fattn-vec-f16.cuh" + +template // D == head size +#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +__launch_bounds__(D, 1) +#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +static __global__ void flash_attn_vec_ext_f16( + const char * __restrict__ Q, + const char * __restrict__ K, + const char * __restrict__ V, + const char * __restrict__ mask, + float * __restrict__ dst, + float2 * __restrict__ dst_meta, + const float scale, + const float max_bias, + const float m0, + const float m1, + const uint32_t n_head_log2, + const int ne00, + const int ne01, + const int ne02, + const int ne03, + const int ne10, + const int ne11, + const int ne12, + const int ne13, + const int ne31, + const int nb31, + const int nb01, + const int nb02, + const int nb03, + const int nb11, + const int nb12, + const int nb13, + const int ne0, + const int ne1, + const int ne2, + const int ne3) { +#if FP16_AVAILABLE + //In this kernel Q, K, V are matrices while i, j, k are matrix indices. + + const int ic0 = (blockIdx.x / parallel_blocks) * ncols; // Index of the Q/QKV column to work on. + const int ip = blockIdx.x % parallel_blocks; // Index in group of blocks running for the same column in parallel. + + const int gqa_ratio = ne02 / ne12; // With grouped query attention there are > 1 Q matrices per K, V matrix. + const float2 * Q_f2 = (const float2 *) (Q + nb02* blockIdx.y + nb01*ic0); + const half2 * K_h2 = (const half2 *) (K + nb12*(blockIdx.y / gqa_ratio)); + const half * V_h = (const half *) (V + nb12*(blockIdx.y / gqa_ratio)); // K and V have same shape + const half * maskh = (const half *) mask + ne11*ic0; + + const int stride_KV = nb11 / sizeof(half); + const int stride_KV2 = nb11 / sizeof(half2); + + half slopeh = __float2half(1.0f); + + // ALiBi + if (max_bias > 0.0f) { + const int h = blockIdx.y; + + const float base = h < n_head_log2 ? m0 : m1; + const int exph = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; + + slopeh = __float2half(powf(base, exph)); + } + + static_assert(D % (2*WARP_SIZE) == 0, "D not divisible by 2*WARP_SIZE == 64."); + constexpr int nwarps = D / WARP_SIZE; + const int tid = WARP_SIZE*threadIdx.y + threadIdx.x; + __builtin_assume(tid < D); + + __shared__ half KQ[ncols*D]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + KQ[j*D + tid] = -HALF_MAX_HALF; + } + half2 * KQ2 = (half2 *) KQ; + + half kqmax[ncols]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + kqmax[j] = -HALF_MAX_HALF; + } + half kqsum[ncols] = {0.0f}; + + __shared__ half kqmax_shared[ncols][WARP_SIZE]; + __shared__ half kqsum_shared[ncols][WARP_SIZE]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + if (threadIdx.y == 0) { + kqmax_shared[j][threadIdx.x] = -HALF_MAX_HALF; + kqsum_shared[j][threadIdx.x] = 0.0f; + } + } + __syncthreads(); + + // Convert Q to half2 and store in registers: + half2 Q_h2[ncols][D/(2*WARP_SIZE)]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { +#pragma unroll + for (int i0 = 0; i0 < D/2; i0 += WARP_SIZE) { + const int i = i0 + threadIdx.x; + + const float2 tmp = Q_f2[j*(nb01/sizeof(float2)) + i]; + Q_h2[j][i0/WARP_SIZE] = make_half2(scale, scale) * make_half2(tmp.x, tmp.y); + } + } + + half2 VKQ[ncols] = {{0.0f, 0.0f}}; + + const int k_start = parallel_blocks == 1 ? 0 : ip*D; + for (int k_VKQ_0 = k_start; k_VKQ_0 < ne11; k_VKQ_0 += parallel_blocks*D) { + // Calculate KQ tile and keep track of new maximum KQ values: + + // For unknown reasons using a half array of size 1 for kqmax_new causes a performance regression, + // see https://github.com/ggerganov/llama.cpp/pull/7061 . + // Therefore this variable is defined twice but only used once (so that the compiler can optimize out the unused variable). + half kqmax_new = kqmax[0]; + half kqmax_new_arr[ncols]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + kqmax_new_arr[j] = kqmax[j]; + } + +#pragma unroll + for (int i_KQ_0 = 0; i_KQ_0 < D; i_KQ_0 += nwarps) { + const int i_KQ = i_KQ_0 + threadIdx.y; + + if ((i_KQ_0 + nwarps > D && i_KQ >= D) || (FATTN_KQ_STRIDE % D != 0 && k_VKQ_0 + i_KQ >= ne11)) { + break; + } + + half2 sum2[ncols] = {{0.0f, 0.0f}}; +#pragma unroll + for (int k_KQ_0 = 0; k_KQ_0 < D/2; k_KQ_0 += WARP_SIZE) { + const int k_KQ = k_KQ_0 + threadIdx.x; + + const half2 K_ik = K_h2[(k_VKQ_0 + i_KQ)*stride_KV2 + k_KQ]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + sum2[j] += K_ik * Q_h2[j][k_KQ_0/WARP_SIZE]; + } + } + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + sum2[j] = warp_reduce_sum(sum2[j]); + half sum = __low2half(sum2[j]) + __high2half(sum2[j]); + sum += mask ? slopeh*maskh[j*ne11 + k_VKQ_0 + i_KQ] : __float2half(0.0f); + + if (ncols == 1) { + kqmax_new = ggml_cuda_hmax(kqmax_new, sum); + } else { + kqmax_new_arr[j] = ggml_cuda_hmax(kqmax_new_arr[j], sum); + } + + if (threadIdx.x == 0) { + KQ[j*D + i_KQ] = sum; + } + } + } + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + half kqmax_new_j = ncols == 1 ? kqmax_new : kqmax_new_arr[j]; + + kqmax_new_j = warp_reduce_max(kqmax_new_j); + if (threadIdx.x == 0) { + kqmax_shared[j][threadIdx.y] = kqmax_new_j; + } + } + + __syncthreads(); + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + half kqmax_new_j = kqmax_shared[j][threadIdx.x]; + kqmax_new_j = warp_reduce_max(kqmax_new_j); + + const half KQ_max_scale = hexp(kqmax[j] - kqmax_new_j); + kqmax[j] = kqmax_new_j; + + const half val = hexp(KQ[j*D + tid] - kqmax[j]); + kqsum[j] = kqsum[j]*KQ_max_scale + val; + KQ[j*D + tid] = val; + + VKQ[j] *= __half2half2(KQ_max_scale); + } + + __syncthreads(); + +#pragma unroll + for (int k0 = 0; k0 < D; k0 += 2) { + if (FATTN_KQ_STRIDE % D != 0 && k_VKQ_0 + k0 >= ne11) { + break; + } + + half2 V_k; + reinterpret_cast(V_k.x) = V_h[(k_VKQ_0 + k0 + 0)*stride_KV + tid]; + reinterpret_cast(V_k.y) = V_h[(k_VKQ_0 + k0 + 1)*stride_KV + tid]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + VKQ[j] += V_k*KQ2[j*(D/2) + k0/2]; + } + } + + __syncthreads(); + } + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + kqsum[j] = warp_reduce_sum(kqsum[j]); + if (threadIdx.x == 0) { + kqsum_shared[j][threadIdx.y] = kqsum[j]; + } + } + + __syncthreads(); + +#pragma unroll + for (int j_VKQ = 0; j_VKQ < ncols; ++j_VKQ) { + kqsum[j_VKQ] = kqsum_shared[j_VKQ][threadIdx.x]; + kqsum[j_VKQ] = warp_reduce_sum(kqsum[j_VKQ]); + + half dst_val = (__low2half(VKQ[j_VKQ]) + __high2half(VKQ[j_VKQ])); + if (parallel_blocks == 1) { + dst_val /= kqsum[j_VKQ]; + } + const int j_dst = (ic0 + j_VKQ)*parallel_blocks + ip; + dst[j_dst*D*gridDim.y + D*blockIdx.y + tid] = dst_val; + } + + if (parallel_blocks != 1 && tid != 0) { +#pragma unroll + for (int j = 0; j < ncols; ++j) { + dst_meta[(ic0 + j)*gridDim.y*parallel_blocks + blockIdx.y*parallel_blocks + ip] = make_float2(kqmax[j], kqsum[j]); + } + } +#else + NO_DEVICE_CODE; +#endif // FP16_AVAILABLE +} + +template void launch_fattn_vec_f16( + const ggml_tensor * Q, const ggml_tensor * K, const ggml_tensor * V, ggml_tensor * KQV, const ggml_tensor * mask, + ggml_cuda_pool & pool, cudaStream_t main_stream +) { + ggml_cuda_pool_alloc dst_tmp(pool); + ggml_cuda_pool_alloc dst_tmp_meta(pool); + + if (parallel_blocks > 1) { + dst_tmp.alloc(parallel_blocks*ggml_nelements(KQV)); + dst_tmp_meta.alloc(parallel_blocks*ggml_nrows(KQV)); + } + + constexpr int nwarps = (D + WARP_SIZE - 1) / WARP_SIZE; + const dim3 block_dim(WARP_SIZE, nwarps, 1); + const dim3 blocks_num(parallel_blocks*((Q->ne[1] + cols_per_block - 1) / cols_per_block), Q->ne[2], Q->ne[3]); + const int shmem = 0; + + float scale = 1.0f; + float max_bias = 0.0f; + + memcpy(&scale, (float *) KQV->op_params + 0, sizeof(float)); + memcpy(&max_bias, (float *) KQV->op_params + 1, sizeof(float)); + + const uint32_t n_head = Q->ne[2]; + const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); + + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + flash_attn_vec_ext_f16 + <<>> ( + (const char *) Q->data, + (const char *) K->data, + (const char *) V->data, + mask ? ((const char *) mask->data) : nullptr, + parallel_blocks == 1 ? (float *) KQV->data : dst_tmp.ptr, dst_tmp_meta.ptr, + scale, max_bias, m0, m1, n_head_log2, + Q->ne[0], Q->ne[1], Q->ne[2], Q->ne[3], + K->ne[0], K->ne[1], K->ne[2], K->ne[3], + mask ? mask->ne[1] : 0, mask ? mask->nb[1] : 0, + Q->nb[1], Q->nb[2], Q->nb[3], + K->nb[1], K->nb[2], K->nb[3], + KQV->ne[0], KQV->ne[1], KQV->ne[2], KQV->ne[3] + ); + CUDA_CHECK(cudaGetLastError()); + + if (parallel_blocks == 1) { + return; + } + + const dim3 block_dim_combine(D, 1, 1); + const dim3 blocks_num_combine(Q->ne[1], blocks_num.y, blocks_num.z); + const int shmem_combine = 0; + + flash_attn_combine_results + <<>> + (dst_tmp.ptr, dst_tmp_meta.ptr, (float *) KQV->data); + CUDA_CHECK(cudaGetLastError()); +} + +void ggml_cuda_flash_attn_ext_vec_f16(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { + const ggml_tensor * Q = dst->src[0]; + const ggml_tensor * K = dst->src[1]; + const ggml_tensor * V = dst->src[2]; + + const ggml_tensor * mask = dst->src[3]; + + ggml_tensor * KQV = dst; + + const int32_t precision = KQV->op_params[2]; + GGML_ASSERT(precision == GGML_PREC_DEFAULT); + + constexpr int cols_per_block = 1; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 256: + launch_fattn_vec_f16<256, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } +} + +void ggml_cuda_flash_attn_ext_vec_f16_no_mma(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { + const ggml_tensor * Q = dst->src[0]; + const ggml_tensor * K = dst->src[1]; + const ggml_tensor * V = dst->src[2]; + + const ggml_tensor * mask = dst->src[3]; + + ggml_tensor * KQV = dst; + + const int32_t precision = KQV->op_params[2]; + GGML_ASSERT(precision == GGML_PREC_DEFAULT); + GGML_ASSERT(Q->ne[0] == 64 || Q->ne[0] == 128 && "FlashAttention without tensor cores only supports head sizes 64 and 128."); + + if (Q->ne[1] == 1) { + constexpr int cols_per_block = 1; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + if (Q->ne[1] == 2) { + constexpr int cols_per_block = 2; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + if (Q->ne[1] <= 4) { + constexpr int cols_per_block = 4; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + if (Q->ne[1] <= 8) { + constexpr int cols_per_block = 8; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + constexpr int cols_per_block = 8; + constexpr int parallel_blocks = 1; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } +} diff --git a/ggml-cuda/fattn-vec-f16.cuh b/ggml-cuda/fattn-vec-f16.cuh new file mode 100644 index 00000000000..c7023610ab2 --- /dev/null +++ b/ggml-cuda/fattn-vec-f16.cuh @@ -0,0 +1,5 @@ +#include "common.cuh" + +void ggml_cuda_flash_attn_ext_vec_f16(ggml_backend_cuda_context & ctx, ggml_tensor * dst); + +void ggml_cuda_flash_attn_ext_vec_f16_no_mma(ggml_backend_cuda_context & ctx, ggml_tensor * dst); diff --git a/ggml-cuda/fattn-vec-f32.cu b/ggml-cuda/fattn-vec-f32.cu new file mode 100644 index 00000000000..40c336ce332 --- /dev/null +++ b/ggml-cuda/fattn-vec-f32.cu @@ -0,0 +1,384 @@ +#include "common.cuh" +#include "fattn-common.cuh" +#include "fattn-vec-f32.cuh" + +template // D == head size +#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +__launch_bounds__(D, 1) +#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) +static __global__ void flash_attn_vec_ext_f32( + const char * __restrict__ Q, + const char * __restrict__ K, + const char * __restrict__ V, + const char * __restrict__ mask, + float * __restrict__ dst, + float2 * __restrict__ dst_meta, + const float scale, + const float max_bias, + const float m0, + const float m1, + const uint32_t n_head_log2, + const int ne00, + const int ne01, + const int ne02, + const int ne03, + const int ne10, + const int ne11, + const int ne12, + const int ne13, + const int ne31, + const int nb31, + const int nb01, + const int nb02, + const int nb03, + const int nb11, + const int nb12, + const int nb13, + const int ne0, + const int ne1, + const int ne2, + const int ne3) { + //In this kernel Q, K, V are matrices while i, j, k are matrix indices. + + const int ic0 = (blockIdx.x / parallel_blocks) * ncols; // Index of the Q/QKV column to work on. + const int ip = blockIdx.x % parallel_blocks; // Index in group of blocks running for the same column in parallel. + + const int gqa_ratio = ne02 / ne12; // With grouped query attention there are > 1 Q matrices per K, V matrix. + const float2 * Q_f2 = (const float2 *) (Q + nb02* blockIdx.y + nb01*ic0); + const half2 * K_h2 = (const half2 *) (K + nb12*(blockIdx.y / gqa_ratio)); + const half * V_h = (const half *) (V + nb12*(blockIdx.y / gqa_ratio)); // K and V have same shape + const half * maskh = (const half *) mask + ne11*ic0; + + const int stride_KV = nb11 / sizeof(half); + const int stride_KV2 = nb11 / sizeof(half2); + + float slope = 1.0f; + + // ALiBi + if (max_bias > 0.0f) { + const int h = blockIdx.y; + + const float base = h < n_head_log2 ? m0 : m1; + const int exph = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; + + slope = powf(base, exph); + } + + static_assert(D % (2*WARP_SIZE) == 0, "D not divisible by 2*WARP_SIZE == 64."); + constexpr int nwarps = D / WARP_SIZE; + const int tid = WARP_SIZE*threadIdx.y + threadIdx.x; + __builtin_assume(tid < D); + + __shared__ float KQ[ncols*D]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + KQ[j*D + tid] = -FLT_MAX/2.0f; + } + + float kqmax[ncols]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + kqmax[j] = -FLT_MAX/2.0f; + } + float kqsum[ncols] = {0.0f}; + + __shared__ float kqmax_shared[ncols][WARP_SIZE]; + __shared__ float kqsum_shared[ncols][WARP_SIZE]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + if (threadIdx.y == 0) { + kqmax_shared[j][threadIdx.x] = -FLT_MAX/2.0f; + kqsum_shared[j][threadIdx.x] = 0.0f; + } + } + __syncthreads(); + + // Convert Q to half2 and store in registers: + float2 Q_h2[ncols][D/(2*WARP_SIZE)]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { +#pragma unroll + for (int i0 = 0; i0 < D/2; i0 += WARP_SIZE) { + const int i = i0 + threadIdx.x; + + Q_h2[j][i0/WARP_SIZE] = Q_f2[j*(nb01/sizeof(float2)) + i]; + Q_h2[j][i0/WARP_SIZE].x *= scale; + Q_h2[j][i0/WARP_SIZE].y *= scale; + } + } + + float VKQ[ncols] = {0.0f}; + + const int k_start = parallel_blocks == 1 ? 0 : ip*D; + for (int k_VKQ_0 = k_start; k_VKQ_0 < ne11; k_VKQ_0 += parallel_blocks*D) { + // Calculate KQ tile and keep track of new maximum KQ values: + + float kqmax_new_arr[ncols]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + kqmax_new_arr[j] = kqmax[j]; + } + +#pragma unroll + for (int i_KQ_0 = 0; i_KQ_0 < D; i_KQ_0 += nwarps) { + const int i_KQ = i_KQ_0 + threadIdx.y; + + if ((i_KQ_0 + nwarps > D && i_KQ >= D) || (FATTN_KQ_STRIDE % D != 0 && k_VKQ_0 + i_KQ >= ne11)) { + break; + } + + float sum[ncols] = {0.0f}; +#pragma unroll + for (int k_KQ_0 = 0; k_KQ_0 < D/2; k_KQ_0 += WARP_SIZE) { + const int k_KQ = k_KQ_0 + threadIdx.x; + + const half2 K_ik = K_h2[(k_VKQ_0 + i_KQ)*stride_KV2 + k_KQ]; +#pragma unroll + for (int j = 0; j < ncols; ++j) { + sum[j] += __low2float(K_ik) * Q_h2[j][k_KQ_0/WARP_SIZE].x; + sum[j] += __high2float(K_ik) * Q_h2[j][k_KQ_0/WARP_SIZE].y; + } + } + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + sum[j] = warp_reduce_sum(sum[j]); + sum[j] += mask ? slope*__half2float(maskh[j*ne11 + k_VKQ_0 + i_KQ]) : 0.0f; + + kqmax_new_arr[j] = fmaxf(kqmax_new_arr[j], sum[j]); + + if (threadIdx.x == 0) { + KQ[j*D + i_KQ] = sum[j]; + } + } + } + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + float kqmax_new_j = kqmax_new_arr[j]; + + kqmax_new_j = warp_reduce_max(kqmax_new_j); + if (threadIdx.x == 0) { + kqmax_shared[j][threadIdx.y] = kqmax_new_j; + } + } + + __syncthreads(); + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + float kqmax_new_j = kqmax_shared[j][threadIdx.x]; + kqmax_new_j = warp_reduce_max(kqmax_new_j); + + const float KQ_max_scale = expf(kqmax[j] - kqmax_new_j); + kqmax[j] = kqmax_new_j; + + const float val = expf(KQ[j*D + tid] - kqmax[j]); + kqsum[j] = kqsum[j]*KQ_max_scale + val; + KQ[j*D + tid] = val; + + VKQ[j] *= KQ_max_scale; + } + + __syncthreads(); + +#pragma unroll + for (int k = 0; k < D; ++k) { + if (FATTN_KQ_STRIDE % D != 0 && k_VKQ_0 + k >= ne11) { + break; + } + + const float V_ki = __half2float(V_h[(k_VKQ_0 + k)*stride_KV + tid]); +#pragma unroll + for (int j = 0; j < ncols; ++j) { + VKQ[j] += V_ki*KQ[j*D + k]; + } + } + + __syncthreads(); + } + +#pragma unroll + for (int j = 0; j < ncols; ++j) { + kqsum[j] = warp_reduce_sum(kqsum[j]); + if (threadIdx.x == 0) { + kqsum_shared[j][threadIdx.y] = kqsum[j]; + } + } + + __syncthreads(); + +#pragma unroll + for (int j_VKQ = 0; j_VKQ < ncols; ++j_VKQ) { + kqsum[j_VKQ] = kqsum_shared[j_VKQ][threadIdx.x]; + kqsum[j_VKQ] = warp_reduce_sum(kqsum[j_VKQ]); + + float dst_val = VKQ[j_VKQ]; + if (parallel_blocks == 1) { + dst_val /= kqsum[j_VKQ]; + } + const int j_dst = (ic0 + j_VKQ)*parallel_blocks + ip; + dst[j_dst*D*gridDim.y + D*blockIdx.y + tid] = dst_val; + } + + if (parallel_blocks != 1 && tid != 0) { +#pragma unroll + for (int j = 0; j < ncols; ++j) { + dst_meta[(ic0 + j)*gridDim.y*parallel_blocks + blockIdx.y*parallel_blocks + ip] = make_float2(kqmax[j], kqsum[j]); + } + } +} + +template void launch_fattn_vec_f32( + const ggml_tensor * Q, const ggml_tensor * K, const ggml_tensor * V, ggml_tensor * KQV, const ggml_tensor * mask, + ggml_cuda_pool & pool, cudaStream_t main_stream +) { + ggml_cuda_pool_alloc dst_tmp(pool); + ggml_cuda_pool_alloc dst_tmp_meta(pool); + + if (parallel_blocks > 1) { + dst_tmp.alloc(parallel_blocks*ggml_nelements(KQV)); + dst_tmp_meta.alloc(parallel_blocks*ggml_nrows(KQV)); + } + + constexpr int nwarps = (D + WARP_SIZE - 1) / WARP_SIZE; + const dim3 block_dim(WARP_SIZE, nwarps, 1); + const dim3 blocks_num(parallel_blocks*((Q->ne[1] + cols_per_block - 1) / cols_per_block), Q->ne[2], Q->ne[3]); + const int shmem = 0; + + float scale = 1.0f; + float max_bias = 0.0f; + + memcpy(&scale, (float *) KQV->op_params + 0, sizeof(float)); + memcpy(&max_bias, (float *) KQV->op_params + 1, sizeof(float)); + + const uint32_t n_head = Q->ne[2]; + const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); + + const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); + const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); + + flash_attn_vec_ext_f32 + <<>> ( + (const char *) Q->data, + (const char *) K->data, + (const char *) V->data, + mask ? ((const char *) mask->data) : nullptr, + parallel_blocks == 1 ? (float *) KQV->data : dst_tmp.ptr, dst_tmp_meta.ptr, + scale, max_bias, m0, m1, n_head_log2, + Q->ne[0], Q->ne[1], Q->ne[2], Q->ne[3], + K->ne[0], K->ne[1], K->ne[2], K->ne[3], + mask ? mask->ne[1] : 0, mask ? mask->nb[1] : 0, + Q->nb[1], Q->nb[2], Q->nb[3], + K->nb[1], K->nb[2], K->nb[3], + KQV->ne[0], KQV->ne[1], KQV->ne[2], KQV->ne[3] + ); + CUDA_CHECK(cudaGetLastError()); + + if (parallel_blocks == 1) { + return; + } + + const dim3 block_dim_combine(D, 1, 1); + const dim3 blocks_num_combine(Q->ne[1], blocks_num.y, blocks_num.z); + const int shmem_combine = 0; + + flash_attn_combine_results + <<>> + (dst_tmp.ptr, dst_tmp_meta.ptr, (float *) KQV->data); + CUDA_CHECK(cudaGetLastError()); +} + +void ggml_cuda_flash_attn_ext_vec_f32(ggml_backend_cuda_context & ctx, ggml_tensor * dst) { + const ggml_tensor * Q = dst->src[0]; + const ggml_tensor * K = dst->src[1]; + const ggml_tensor * V = dst->src[2]; + + const ggml_tensor * mask = dst->src[3]; + + ggml_tensor * KQV = dst; + + GGML_ASSERT(Q->ne[0] == 64 || Q->ne[0] == 128 && "FlashAttention without tensor cores only supports head sizes 64 and 128."); + + if (Q->ne[1] == 1) { + constexpr int cols_per_block = 1; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f32< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f32<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + if (Q->ne[1] == 2) { + constexpr int cols_per_block = 2; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f32< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f32<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + if (Q->ne[1] <= 4) { + constexpr int cols_per_block = 4; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f32< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f32<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + if (Q->ne[1] <= 8) { + constexpr int cols_per_block = 8; + constexpr int parallel_blocks = 4; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f32< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f32<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } + return; + } + + constexpr int cols_per_block = 8; + constexpr int parallel_blocks = 1; + switch (Q->ne[0]) { + case 64: + launch_fattn_vec_f32< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + case 128: + launch_fattn_vec_f32<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); + break; + default: + GGML_ASSERT(false); + break; + } +} diff --git a/ggml-cuda/fattn-vec-f32.cuh b/ggml-cuda/fattn-vec-f32.cuh new file mode 100644 index 00000000000..614d54ae392 --- /dev/null +++ b/ggml-cuda/fattn-vec-f32.cuh @@ -0,0 +1,3 @@ +#include "common.cuh" + +void ggml_cuda_flash_attn_ext_vec_f32(ggml_backend_cuda_context & ctx, ggml_tensor * dst); diff --git a/ggml-cuda/fattn.cu b/ggml-cuda/fattn.cu index ac5d6672b30..419f8e752a7 100644 --- a/ggml-cuda/fattn.cu +++ b/ggml-cuda/fattn.cu @@ -1,4 +1,7 @@ #include "common.cuh" +#include "fattn-common.cuh" +#include "fattn-vec-f16.cuh" +#include "fattn-vec-f32.cuh" #include "fattn.cuh" #include @@ -7,251 +10,6 @@ #include #endif -#define FATTN_KQ_STRIDE 256 -#define HALF_MAX_HALF __float2half(65504.0f/2) // Use neg. of this instead of -INFINITY to initialize KQ max vals to avoid NaN upon subtraction. -#define SOFTMAX_FTZ_THRESHOLD -20.0f // Softmax exp. of values smaller than this are flushed to zero to avoid NaNs. - -template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) -__launch_bounds__(D, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) -static __global__ void flash_attn_vec_ext_f16( - const char * __restrict__ Q, - const char * __restrict__ K, - const char * __restrict__ V, - const char * __restrict__ mask, - float * __restrict__ dst, - float2 * __restrict__ dst_meta, - const float scale, - const float max_bias, - const float m0, - const float m1, - const uint32_t n_head_log2, - const int ne00, - const int ne01, - const int ne02, - const int ne03, - const int ne10, - const int ne11, - const int ne12, - const int ne13, - const int ne31, - const int nb31, - const int nb01, - const int nb02, - const int nb03, - const int nb11, - const int nb12, - const int nb13, - const int ne0, - const int ne1, - const int ne2, - const int ne3) { -#if FP16_AVAILABLE - //In this kernel Q, K, V are matrices while i, j, k are matrix indices. - - const int ic0 = (blockIdx.x / parallel_blocks) * ncols; // Index of the Q/QKV column to work on. - const int ip = blockIdx.x % parallel_blocks; // Index in group of blocks running for the same column in parallel. - - const int gqa_ratio = ne02 / ne12; // With grouped query attention there are > 1 Q matrices per K, V matrix. - const float2 * Q_f2 = (const float2 *) (Q + nb02* blockIdx.y + nb01*ic0); - const half2 * K_h2 = (const half2 *) (K + nb12*(blockIdx.y / gqa_ratio)); - const half * V_h = (const half *) (V + nb12*(blockIdx.y / gqa_ratio)); // K and V have same shape - const half * maskh = (const half *) mask + ne11*ic0; - - const int stride_KV = nb11 / sizeof(half); - const int stride_KV2 = nb11 / sizeof(half2); - - half slopeh = __float2half(1.0f); - - // ALiBi - if (max_bias > 0.0f) { - const int h = blockIdx.y; - - const float base = h < n_head_log2 ? m0 : m1; - const int exph = h < n_head_log2 ? h + 1 : 2*(h - n_head_log2) + 1; - - slopeh = __float2half(powf(base, exph)); - } - - static_assert(D % (2*WARP_SIZE) == 0, "D not divisible by 2*WARP_SIZE == 64."); - constexpr int nwarps = D / WARP_SIZE; - const int tid = WARP_SIZE*threadIdx.y + threadIdx.x; - __builtin_assume(tid < D); - - __shared__ half KQ[ncols*D]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { - KQ[j*D + tid] = -HALF_MAX_HALF; - } - half2 * KQ2 = (half2 *) KQ; - - half kqmax[ncols]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { - kqmax[j] = -HALF_MAX_HALF; - } - half kqsum[ncols] = {0.0f}; - - __shared__ half kqmax_shared[ncols][WARP_SIZE]; - __shared__ half kqsum_shared[ncols][WARP_SIZE]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { - if (threadIdx.y == 0) { - kqmax_shared[j][threadIdx.x] = -HALF_MAX_HALF; - kqsum_shared[j][threadIdx.x] = 0.0f; - } - } - __syncthreads(); - - // Convert Q to half2 and store in registers: - half2 Q_h2[ncols][D/(2*WARP_SIZE)]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { -#pragma unroll - for (int i0 = 0; i0 < D/2; i0 += WARP_SIZE) { - const int i = i0 + threadIdx.x; - - const float2 tmp = Q_f2[j*(nb01/sizeof(float2)) + i]; - Q_h2[j][i0/WARP_SIZE] = make_half2(scale, scale) * make_half2(tmp.x, tmp.y); - } - } - - half2 VKQ[ncols] = {{0.0f, 0.0f}}; - - const int k_start = parallel_blocks == 1 ? 0 : ip*D; - for (int k_VKQ_0 = k_start; k_VKQ_0 < ne11; k_VKQ_0 += parallel_blocks*D) { - // Calculate KQ tile and keep track of new maximum KQ values: - - // For unknown reasons using a half array of size 1 for kqmax_new causes a performance regression, - // see https://github.com/ggerganov/llama.cpp/pull/7061 . - // Therefore this variable is defined twice but only used once (so that the compiler can optimize out the unused variable). - half kqmax_new = kqmax[0]; - half kqmax_new_arr[ncols]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { - kqmax_new_arr[j] = kqmax[j]; - } - -#pragma unroll - for (int i_KQ_0 = 0; i_KQ_0 < D; i_KQ_0 += nwarps) { - const int i_KQ = i_KQ_0 + threadIdx.y; - - if ((i_KQ_0 + nwarps > D && i_KQ >= D) || (FATTN_KQ_STRIDE % D != 0 && k_VKQ_0 + i_KQ >= ne11)) { - break; - } - - half2 sum2[ncols] = {{0.0f, 0.0f}}; -#pragma unroll - for (int k_KQ_0 = 0; k_KQ_0 < D/2; k_KQ_0 += WARP_SIZE) { - const int k_KQ = k_KQ_0 + threadIdx.x; - - const half2 K_ik = K_h2[(k_VKQ_0 + i_KQ)*stride_KV2 + k_KQ]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { - sum2[j] += K_ik * Q_h2[j][k_KQ_0/WARP_SIZE]; - } - } - -#pragma unroll - for (int j = 0; j < ncols; ++j) { - sum2[j] = warp_reduce_sum(sum2[j]); - half sum = __low2half(sum2[j]) + __high2half(sum2[j]); - sum += mask ? slopeh*maskh[j*ne11 + k_VKQ_0 + i_KQ] : __float2half(0.0f); - - if (ncols == 1) { - kqmax_new = ggml_cuda_hmax(kqmax_new, sum); - } else { - kqmax_new_arr[j] = ggml_cuda_hmax(kqmax_new_arr[j], sum); - } - - if (threadIdx.x == 0) { - KQ[j*D + i_KQ] = sum; - } - } - } - -#pragma unroll - for (int j = 0; j < ncols; ++j) { - half kqmax_new_j = ncols == 1 ? kqmax_new : kqmax_new_arr[j]; - - kqmax_new_j = warp_reduce_max(kqmax_new_j); - if (threadIdx.x == 0) { - kqmax_shared[j][threadIdx.y] = kqmax_new_j; - } - } - - __syncthreads(); - -#pragma unroll - for (int j = 0; j < ncols; ++j) { - half kqmax_new_j = kqmax_shared[j][threadIdx.x]; - kqmax_new_j = warp_reduce_max(kqmax_new_j); - - const half KQ_max_scale = hexp(kqmax[j] - kqmax_new_j); - kqmax[j] = kqmax_new_j; - - const half val = hexp(KQ[j*D + tid] - kqmax[j]); - kqsum[j] = kqsum[j]*KQ_max_scale + val; - KQ[j*D + tid] = val; - - VKQ[j] *= __half2half2(KQ_max_scale); - } - - __syncthreads(); - -#pragma unroll - for (int k0 = 0; k0 < D; k0 += 2) { - if (FATTN_KQ_STRIDE % D != 0 && k_VKQ_0 + k0 >= ne11) { - break; - } - - half2 V_k; - reinterpret_cast(V_k.x) = V_h[(k_VKQ_0 + k0 + 0)*stride_KV + tid]; - reinterpret_cast(V_k.y) = V_h[(k_VKQ_0 + k0 + 1)*stride_KV + tid]; -#pragma unroll - for (int j = 0; j < ncols; ++j) { - VKQ[j] += V_k*KQ2[j*(D/2) + k0/2]; - } - } - - __syncthreads(); - } - -#pragma unroll - for (int j = 0; j < ncols; ++j) { - kqsum[j] = warp_reduce_sum(kqsum[j]); - if (threadIdx.x == 0) { - kqsum_shared[j][threadIdx.y] = kqsum[j]; - } - } - - __syncthreads(); - -#pragma unroll - for (int j_VKQ = 0; j_VKQ < ncols; ++j_VKQ) { - kqsum[j_VKQ] = kqsum_shared[j_VKQ][threadIdx.x]; - kqsum[j_VKQ] = warp_reduce_sum(kqsum[j_VKQ]); - - half dst_val = (__low2half(VKQ[j_VKQ]) + __high2half(VKQ[j_VKQ])); - if (parallel_blocks == 1) { - dst_val /= kqsum[j_VKQ]; - } - const int j_dst = (ic0 + j_VKQ)*parallel_blocks + ip; - dst[j_dst*D*gridDim.y + D*blockIdx.y + tid] = dst_val; - } - - if (parallel_blocks != 1 && tid != 0) { -#pragma unroll - for (int j = 0; j < ncols; ++j) { - dst_meta[(ic0 + j)*gridDim.y*parallel_blocks + blockIdx.y*parallel_blocks + ip] = make_float2(kqmax[j], kqsum[j]); - } - } -#else - NO_DEVICE_CODE; -#endif // FP16_AVAILABLE -} - // D == head size, VKQ_stride == num VKQ rows calculated in parallel: template #if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) @@ -655,54 +413,6 @@ static __global__ void flash_attn_ext_f16( #endif // FP16_MMA_AVAILABLE } -template // D == head size -#if !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) -__launch_bounds__(D, 1) -#endif // !(defined(GGML_USE_HIPBLAS) && defined(__HIP_PLATFORM_AMD__)) -static __global__ void flash_attn_combine_results( - const float * __restrict__ VKQ_parts, - const float2 * __restrict__ VKQ_meta, - float * __restrict__ dst) { -#if FP16_AVAILABLE - VKQ_parts += parallel_blocks*D * gridDim.y*blockIdx.x; - VKQ_meta += parallel_blocks * gridDim.y*blockIdx.x; - dst += D * gridDim.y*blockIdx.x; - - const int tid = threadIdx.x; - __builtin_assume(tid < D); - - __shared__ float2 meta[parallel_blocks]; - if (tid < 2*parallel_blocks) { - ((float *) meta)[threadIdx.x] = ((const float *)VKQ_meta) [blockIdx.y*(2*parallel_blocks) + tid]; - } - - __syncthreads(); - - float kqmax = meta[0].x; -#pragma unroll - for (int l = 1; l < parallel_blocks; ++l) { - kqmax = max(kqmax, meta[l].x); - } - - float VKQ_numerator = 0.0f; - float VKQ_denominator = 0.0f; -#pragma unroll - for (int l = 0; l < parallel_blocks; ++l) { - const float diff = meta[l].x - kqmax; - const float KQ_max_scale = expf(diff); - const uint32_t ftz_mask = 0xFFFFFFFF * (diff > SOFTMAX_FTZ_THRESHOLD); - *((uint32_t *) &KQ_max_scale) &= ftz_mask; - - VKQ_numerator += KQ_max_scale * VKQ_parts[l*gridDim.y*D + blockIdx.y*D + tid]; - VKQ_denominator += KQ_max_scale * meta[l].y; - } - - dst[blockIdx.y*D + tid] = VKQ_numerator / VKQ_denominator; -#else - NO_DEVICE_CODE; -#endif // FP16_AVAILABLE -} - constexpr int get_max_power_of_2(int x) { return x % 2 == 0 ? 2*get_max_power_of_2(x/2) : 1; } @@ -727,66 +437,6 @@ static_assert(get_VKQ_stride( 80, 1, 16) == 16, "Test failed."); static_assert(get_VKQ_stride( 80, 2, 16) == 16, "Test failed."); static_assert(get_VKQ_stride( 80, 4, 16) == 16, "Test failed."); -template void launch_fattn_vec_f16( - const ggml_tensor * Q, const ggml_tensor * K, const ggml_tensor * V, ggml_tensor * KQV, const ggml_tensor * mask, - ggml_cuda_pool & pool, cudaStream_t main_stream -) { - ggml_cuda_pool_alloc dst_tmp(pool); - ggml_cuda_pool_alloc dst_tmp_meta(pool); - - if (parallel_blocks > 1) { - dst_tmp.alloc(parallel_blocks*ggml_nelements(KQV)); - dst_tmp_meta.alloc(parallel_blocks*ggml_nrows(KQV)); - } - - constexpr int nwarps = (D + WARP_SIZE - 1) / WARP_SIZE; - const dim3 block_dim(WARP_SIZE, nwarps, 1); - const dim3 blocks_num(parallel_blocks*((Q->ne[1] + cols_per_block - 1) / cols_per_block), Q->ne[2], Q->ne[3]); - const int shmem = 0; - - float scale = 1.0f; - float max_bias = 0.0f; - - memcpy(&scale, (float *) KQV->op_params + 0, sizeof(float)); - memcpy(&max_bias, (float *) KQV->op_params + 1, sizeof(float)); - - const uint32_t n_head = Q->ne[2]; - const uint32_t n_head_log2 = 1u << (uint32_t) floorf(log2f((float) n_head)); - - const float m0 = powf(2.0f, -(max_bias ) / n_head_log2); - const float m1 = powf(2.0f, -(max_bias / 2.0f) / n_head_log2); - - flash_attn_vec_ext_f16 - <<>> ( - (const char *) Q->data, - (const char *) K->data, - (const char *) V->data, - mask ? ((const char *) mask->data) : nullptr, - parallel_blocks == 1 ? (float *) KQV->data : dst_tmp.ptr, dst_tmp_meta.ptr, - scale, max_bias, m0, m1, n_head_log2, - Q->ne[0], Q->ne[1], Q->ne[2], Q->ne[3], - K->ne[0], K->ne[1], K->ne[2], K->ne[3], - mask ? mask->ne[1] : 0, mask ? mask->nb[1] : 0, - Q->nb[1], Q->nb[2], Q->nb[3], - K->nb[1], K->nb[2], K->nb[3], - KQV->ne[0], KQV->ne[1], KQV->ne[2], KQV->ne[3] - ); - CUDA_CHECK(cudaGetLastError()); - - if (parallel_blocks == 1) { - return; - } - - const dim3 block_dim_combine(D, 1, 1); - const dim3 blocks_num_combine(Q->ne[1], blocks_num.y, blocks_num.z); - const int shmem_combine = 0; - - flash_attn_combine_results - <<>> - (dst_tmp.ptr, dst_tmp_meta.ptr, (float *) KQV->data); - CUDA_CHECK(cudaGetLastError()); -} - template void launch_fattn_f16_impl( const ggml_tensor * Q, const ggml_tensor * K, const ggml_tensor * V, ggml_tensor * KQV, const ggml_tensor * mask, ggml_cuda_pool & pool, cudaStream_t main_stream @@ -891,95 +541,22 @@ void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst const int32_t precision = KQV->op_params[2]; - if (!fp16_mma_available(cc)) { - GGML_ASSERT(precision == GGML_PREC_DEFAULT); - GGML_ASSERT(Q->ne[0] == 64 || Q->ne[0] == 128 && "FlashAttention without tensor cores only supports head sizes 64 and 128."); - - if (Q->ne[1] == 1) { - constexpr int cols_per_block = 1; - constexpr int parallel_blocks = 4; - switch (Q->ne[0]) { - case 64: - launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 128: - launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - default: - GGML_ASSERT(false); - break; - } - return; - } - - if (Q->ne[1] == 2) { - constexpr int cols_per_block = 2; - constexpr int parallel_blocks = 4; - switch (Q->ne[0]) { - case 64: - launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 128: - launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - default: - GGML_ASSERT(false); - break; - } - return; - } - - if (Q->ne[1] <= 4) { - constexpr int cols_per_block = 4; - constexpr int parallel_blocks = 4; - switch (Q->ne[0]) { - case 64: - launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 128: - launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - default: - GGML_ASSERT(false); - break; - } - return; - } - - if (Q->ne[1] <= 8) { - constexpr int cols_per_block = 8; - constexpr int parallel_blocks = 4; - switch (Q->ne[0]) { - case 64: - launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 128: - launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - default: - GGML_ASSERT(false); - break; - } - return; - } + if (!fast_fp16_available(cc)) { + ggml_cuda_flash_attn_ext_vec_f32(ctx, dst); + return; + } - constexpr int cols_per_block = 8; - constexpr int parallel_blocks = 1; - switch (Q->ne[0]) { - case 64: - launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 128: - launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - default: - GGML_ASSERT(false); - break; - } + if (!fp16_mma_available(cc)) { + ggml_cuda_flash_attn_ext_vec_f16_no_mma(ctx, dst); return; } if (precision != GGML_PREC_DEFAULT) { + if (Q->ne[1] == 1 && (Q->ne[0] == 64 || Q->ne[0] == 128)) { + ggml_cuda_flash_attn_ext_vec_f32(ctx, dst); + return; + } + if (Q->ne[1] <= 32 || Q->ne[0] > 128) { constexpr int cols_per_block = 16; constexpr int nwarps = 4; @@ -1037,22 +614,7 @@ void ggml_cuda_flash_attn_ext(ggml_backend_cuda_context & ctx, ggml_tensor * dst } if (Q->ne[1] == 1 && Q->ne[0] % (2*WARP_SIZE) == 0) { - constexpr int cols_per_block = 1; - constexpr int parallel_blocks = 4; - switch (Q->ne[0]) { - case 64: - launch_fattn_vec_f16< 64, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 128: - launch_fattn_vec_f16<128, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - case 256: - launch_fattn_vec_f16<256, cols_per_block, parallel_blocks>(Q, K, V, KQV, mask, ctx.pool(), ctx.stream()); - break; - default: - GGML_ASSERT(false); - break; - } + ggml_cuda_flash_attn_ext_vec_f16(ctx, dst); return; } From 8e7c22fbdbc27c1c72abd192720e330e7f6361a9 Mon Sep 17 00:00:00 2001 From: Neo Zhang <14088817+arthw@users.noreply.github.com> Date: Mon, 13 May 2024 18:11:26 +0800 Subject: [PATCH 03/17] rm wait() (llama/7233) --- ggml-sycl.cpp | 25 +------------------------ 1 file changed, 1 insertion(+), 24 deletions(-) diff --git a/ggml-sycl.cpp b/ggml-sycl.cpp index e93d2af631c..724070eb910 100644 --- a/ggml-sycl.cpp +++ b/ggml-sycl.cpp @@ -15564,26 +15564,6 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0, const int64_t r2 = ne12/ne02; const int64_t r3 = ne13/ne03; -#if 0 - // use syclGemmEx - { - for (int i13 = 0; i13 < ne13; ++i13) { - for (int i12 = 0; i12 < ne12; ++i12) { - int i03 = i13 / r3; - int i02 = i12 / r2; - - SYCL_CHECK( - syclGemmEx(g_sycl_handles[g_main_device], CUBLAS_OP_T, CUBLAS_OP_N, - ne01, ne11, ne10, - alpha, (const char *) src0_as_f16 + i02*src0->nb[2] + i03*src0->nb[3] , SYCL_R_16F, nb01/sizeof(half), - (const char *) src1_as_f16 + i12*src1->nb[2]/2 + i13*src1->nb[3]/2, SYCL_R_16F, nb11/sizeof(float), - beta, ( char *) dst_t + i12*nbd2 + i13*nbd3, cu_data_type, ne01, - cu_compute_type, - CUBLAS_GEMM_DEFAULT_TENSOR_OP)); - } - } - } -#else if (r2 == 1 && r3 == 1 && src0->nb[2]*src0->ne[2] == src0->nb[3] && src1->nb[2]*src1->ne[2] == src1->nb[3]) { // there is no broadcast and src0, src1 are contiguous across dims 2, 3 SYCL_CHECK(CHECK_TRY_ERROR(dpct::gemm_batch( @@ -15595,7 +15575,6 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0, nb11 / nb10, nb12 / nb10, beta, (char *)dst_t, cu_data_type, ne01, nb2 / nb0, ne12 * ne13, cu_compute_type))); - g_sycl_handles[g_main_device]->wait(); } else { const int ne23 = ne12*ne13; @@ -15626,7 +15605,7 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0, nb02, nb03, nb12_scaled, nb13_scaled, nbd2, nbd3, r2, r3, item_ct1); }); - }).wait(); + }); } SYCL_CHECK(CHECK_TRY_ERROR(dpct::gemm_batch( *g_sycl_handles[g_main_device], oneapi::mkl::transpose::trans, @@ -15637,9 +15616,7 @@ static void ggml_sycl_mul_mat_batched_sycl(const ggml_tensor *src0, dpct::library_data_t::real_half, nb11 / nb10, beta, (void **)(ptrs_dst.get() + 0 * ne23), cu_data_type, ne01, ne23, cu_compute_type))); - g_sycl_handles[g_main_device]->wait(); } -#endif if (no_mixed_dtypes) { const to_fp32_sycl_t to_fp32_sycl = ggml_get_to_fp32_sycl(GGML_TYPE_F16); From c451080c8b0e2080f2ca887047ef381b94523e14 Mon Sep 17 00:00:00 2001 From: Radoslav Gerganov Date: Tue, 14 May 2024 14:27:19 +0300 Subject: [PATCH 04/17] ggml : add RPC backend (llama/6829) * ggml : add RPC backend The RPC backend proxies all operations to a remote server which runs a regular backend (CPU, CUDA, Metal, etc). * set TCP_NODELAY * add CI workflows * Address review comments * fix warning * implement llama_max_devices() for RPC * Address review comments * Address review comments * wrap sockfd into a struct * implement get_alignment and get_max_size * add get_device_memory * fix warning * win32 support * add README * readme : trim trailing whitespace * Address review comments * win32 fix * Address review comments * fix compile warnings on macos --- ggml-rpc.cpp | 1023 ++++++++++++++++++++++++++++++++++++++++++++++++++ ggml-rpc.h | 24 ++ 2 files changed, 1047 insertions(+) create mode 100644 ggml-rpc.cpp create mode 100644 ggml-rpc.h diff --git a/ggml-rpc.cpp b/ggml-rpc.cpp new file mode 100644 index 00000000000..efeacb29767 --- /dev/null +++ b/ggml-rpc.cpp @@ -0,0 +1,1023 @@ +#include "ggml-rpc.h" +#include "ggml.h" +#include "ggml-backend-impl.h" + +#include +#include +#include +#include +#include +#include +#ifdef _WIN32 +# define WIN32_LEAN_AND_MEAN +# ifndef NOMINMAX +# define NOMINMAX +# endif +# include +# include +#else +# include +# include +# include +# include +# include +# include +# include +#endif +#include + +#define UNUSED GGML_UNUSED + +#define GGML_DEBUG 1 +#if (GGML_DEBUG >= 1) +#define GGML_PRINT_DEBUG(...) printf(__VA_ARGS__) +#else +#define GGML_PRINT_DEBUG(...) +#endif + +#ifdef _WIN32 +typedef SOCKET sockfd_t; +using ssize_t = __int64; +#else +typedef int sockfd_t; +#endif + +// cross-platform socket +struct socket_t { + sockfd_t fd; + socket_t(sockfd_t fd) : fd(fd) {} + ~socket_t() { +#ifdef _WIN32 + closesocket(this->fd); +#else + close(this->fd); +#endif + } +}; + +// ggml_tensor is serialized into rpc_tensor +struct rpc_tensor { + uint64_t id; + uint32_t type; + uint64_t buffer; + uint32_t ne[GGML_MAX_DIMS]; + uint32_t nb[GGML_MAX_DIMS]; + uint32_t op; + int32_t op_params[GGML_MAX_OP_PARAMS / sizeof(int32_t)]; + int32_t flags; + uint64_t src[GGML_MAX_SRC]; + uint64_t view_src; + uint64_t view_offs; + uint64_t data; + char name[GGML_MAX_NAME]; +}; + +// RPC commands +enum rpc_cmd { + ALLOC_BUFFER = 0, + GET_ALIGNMENT, + GET_MAX_SIZE, + BUFFER_GET_BASE, + FREE_BUFFER, + BUFFER_CLEAR, + SET_TENSOR, + GET_TENSOR, + COPY_TENSOR, + GRAPH_COMPUTE, + GET_DEVICE_MEMORY, +}; + +// RPC data structures + +static ggml_guid_t ggml_backend_rpc_guid() { + static ggml_guid guid = {0x99, 0x68, 0x5b, 0x6c, 0xd2, 0x83, 0x3d, 0x24, 0x25, 0x36, 0x72, 0xe1, 0x5b, 0x0e, 0x14, 0x03}; + return &guid; +} + +struct ggml_backend_rpc_buffer_type_context { + std::shared_ptr sock; + std::string name; + size_t alignment; + size_t max_size; +}; + +struct ggml_backend_rpc_context { + std::string endpoint; + std::string name; + std::shared_ptr sock; + ggml_backend_buffer_type_t buft; +}; + +struct ggml_backend_rpc_buffer_context { + std::shared_ptr sock; + std::unordered_map base_cache; + uint64_t remote_ptr; + std::string name; +}; + +// RPC helper functions + +static std::shared_ptr make_socket(sockfd_t fd) { +#ifdef _WIN32 + if (fd == INVALID_SOCKET) { + return nullptr; + } +#else + if (fd < 0) { + return nullptr; + } +#endif + return std::make_shared(fd); +} + +static bool set_no_delay(sockfd_t sockfd) { + int flag = 1; + // set TCP_NODELAY to disable Nagle's algorithm + int ret = setsockopt(sockfd, IPPROTO_TCP, TCP_NODELAY, (char *)&flag, sizeof(int)); + return ret >= 0; +} + +static std::shared_ptr socket_connect(const char * host, int port) { + struct sockaddr_in addr; + auto sockfd = socket(AF_INET, SOCK_STREAM, 0); + auto sock_ptr = make_socket(sockfd); + if (sock_ptr == nullptr) { + return nullptr; + } + if (!set_no_delay(sockfd)) { + fprintf(stderr, "Failed to set TCP_NODELAY\n"); + return nullptr; + } + addr.sin_family = AF_INET; + addr.sin_port = htons(port); + struct hostent * server = gethostbyname(host); + if (server == NULL) { + fprintf(stderr, "Cannot resolve host '%s'\n", host); + return nullptr; + } + memcpy(&addr.sin_addr.s_addr, server->h_addr, server->h_length); + if (connect(sock_ptr->fd, (struct sockaddr *)&addr, sizeof(addr)) < 0) { + return nullptr; + } + return sock_ptr; +} + +static std::shared_ptr socket_accept(sockfd_t srv_sockfd) { + auto client_socket_fd = accept(srv_sockfd, NULL, NULL); + auto client_socket = make_socket(client_socket_fd); + if (client_socket == nullptr) { + return nullptr; + } + if (!set_no_delay(client_socket_fd)) { + fprintf(stderr, "Failed to set TCP_NODELAY\n"); + return nullptr; + } + return client_socket; +} + +static std::shared_ptr create_server_socket(const char * host, int port) { + auto sockfd = socket(AF_INET, SOCK_STREAM, 0); + auto sock = make_socket(sockfd); + if (sock == nullptr) { + return nullptr; + } + + struct sockaddr_in serv_addr; + serv_addr.sin_family = AF_INET; + serv_addr.sin_addr.s_addr = inet_addr(host); + serv_addr.sin_port = htons(port); + + if (bind(sockfd, (struct sockaddr *) &serv_addr, sizeof(serv_addr)) < 0) { + return nullptr; + } + if (listen(sockfd, 1) < 0) { + return nullptr; + } + return sock; +} + +static bool send_data(sockfd_t sockfd, const void * data, size_t size) { + size_t bytes_sent = 0; + while (bytes_sent < size) { + ssize_t n = send(sockfd, (const char *)data + bytes_sent, size - bytes_sent, 0); + if (n < 0) { + return false; + } + bytes_sent += n; + } + return true; +} + +static bool recv_data(sockfd_t sockfd, void * data, size_t size) { + size_t bytes_recv = 0; + while (bytes_recv < size) { + ssize_t n = recv(sockfd, (char *)data + bytes_recv, size - bytes_recv, 0); + if (n <= 0) { + return false; + } + bytes_recv += n; + } + return true; +} + +static bool parse_endpoint(const char * endpoint, std::string & host, int & port) { + std::string str(endpoint); + size_t pos = str.find(':'); + if (pos == std::string::npos) { + return false; + } + host = str.substr(0, pos); + port = std::stoi(str.substr(pos + 1)); + return true; +} + +// RPC request : | rpc_cmd (1 byte) | request_size (8 bytes) | request_data (request_size bytes) | +// RPC response: | response_size (8 bytes) | response_data (response_size bytes) | +static bool send_rpc_cmd(const std::shared_ptr & sock, enum rpc_cmd cmd, const std::vector & input, std::vector & output) { + uint8_t cmd_byte = cmd; + if (!send_data(sock->fd, &cmd_byte, sizeof(cmd_byte))) { + return false; + } + uint64_t input_size = input.size(); + if (!send_data(sock->fd, &input_size, sizeof(input_size))) { + return false; + } + if (!send_data(sock->fd, input.data(), input.size())) { + return false; + } + uint64_t output_size; + if (!recv_data(sock->fd, &output_size, sizeof(output_size))) { + return false; + } + if (output_size == 0) { + output.clear(); + return true; + } + output.resize(output_size); + if (!recv_data(sock->fd, output.data(), output_size)) { + return false; + } + return true; +} + +// RPC client-side implementation + +GGML_CALL static const char * ggml_backend_rpc_buffer_get_name(ggml_backend_buffer_t buffer) { + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + return ctx->name.c_str(); +} + +GGML_CALL static void ggml_backend_rpc_buffer_free_buffer(ggml_backend_buffer_t buffer) { + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + // input serialization format: | remote_ptr (8 bytes) | + std::vector input(sizeof(uint64_t), 0); + uint64_t remote_ptr = ctx->remote_ptr; + memcpy(input.data(), &remote_ptr, sizeof(remote_ptr)); + std::vector output; + bool status = send_rpc_cmd(ctx->sock, FREE_BUFFER, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.empty()); + delete ctx; +} + +GGML_CALL static void * ggml_backend_rpc_buffer_get_base(ggml_backend_buffer_t buffer) { + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + if (ctx->base_cache.find(buffer) != ctx->base_cache.end()) { + return ctx->base_cache[buffer]; + } + // input serialization format: | remote_ptr (8 bytes) | + std::vector input(sizeof(uint64_t), 0); + uint64_t remote_ptr = ctx->remote_ptr; + memcpy(input.data(), &remote_ptr, sizeof(remote_ptr)); + std::vector output; + bool status = send_rpc_cmd(ctx->sock, BUFFER_GET_BASE, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == sizeof(uint64_t)); + // output serialization format: | base_ptr (8 bytes) | + uint64_t base_ptr; + memcpy(&base_ptr, output.data(), sizeof(base_ptr)); + void * base = reinterpret_cast(base_ptr); + ctx->base_cache[buffer] = base; + return base; +} + +static rpc_tensor serialize_tensor(const ggml_tensor * tensor) { + rpc_tensor result; + result.id = reinterpret_cast(tensor); + result.type = tensor->type; + if (tensor->buffer) { + ggml_backend_buffer_t buffer = tensor->buffer; + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + result.buffer = ctx->remote_ptr; + } else { + result.buffer = 0; + } + for (uint32_t i = 0; i < GGML_MAX_DIMS; i++) { + result.ne[i] = tensor->ne[i]; + result.nb[i] = tensor->nb[i]; + } + result.op = tensor->op; + for (uint32_t i = 0; i < GGML_MAX_OP_PARAMS / sizeof(int32_t); i++) { + result.op_params[i] = tensor->op_params[i]; + } + result.flags = tensor->flags; + for (uint32_t i = 0; i < GGML_MAX_SRC; i++) { + result.src[i] = reinterpret_cast(tensor->src[i]); + } + result.view_src = reinterpret_cast(tensor->view_src); + result.view_offs = tensor->view_offs; + result.data = reinterpret_cast(tensor->data); + snprintf(result.name, GGML_MAX_NAME, "%s", tensor->name); + return result; +} + +static ggml_tensor * deserialize_tensor(struct ggml_context * ctx, const rpc_tensor * tensor) { + ggml_tensor * result = ggml_new_tensor_4d(ctx, (ggml_type) tensor->type, + tensor->ne[0], tensor->ne[1], tensor->ne[2], tensor->ne[3]); + for (uint32_t i = 0; i < GGML_MAX_DIMS; i++) { + result->nb[i] = tensor->nb[i]; + } + result->buffer = reinterpret_cast(tensor->buffer); + result->op = (ggml_op) tensor->op; + for (uint32_t i = 0; i < GGML_MAX_OP_PARAMS / sizeof(int32_t); i++) { + result->op_params[i] = tensor->op_params[i]; + } + result->flags = tensor->flags; + result->data = reinterpret_cast(tensor->data); + ggml_set_name(result, tensor->name); + return result; +} + +GGML_CALL static void ggml_backend_rpc_buffer_init_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor) { + UNUSED(buffer); + if (ggml_is_quantized(tensor->type)) { + // TODO: this check is due to MATRIX_ROW_PADDING in CUDA and should be generalized + GGML_ASSERT(tensor->ne[0] % 512 == 0 && "unsupported quantized tensor"); + } +} + +GGML_CALL static void ggml_backend_rpc_buffer_set_tensor(ggml_backend_buffer_t buffer, ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + // input serialization format: | rpc_tensor | offset (8 bytes) | data (size bytes) | + size_t input_size = sizeof(rpc_tensor) + sizeof(uint64_t) + size; + std::vector input(input_size, 0); + rpc_tensor rpc_tensor = serialize_tensor(tensor); + memcpy(input.data(), &rpc_tensor, sizeof(rpc_tensor)); + memcpy(input.data() + sizeof(rpc_tensor), &offset, sizeof(offset)); + memcpy(input.data() + sizeof(rpc_tensor) + sizeof(offset), data, size); + std::vector output; + bool status = send_rpc_cmd(ctx->sock, SET_TENSOR, input, output); + GGML_ASSERT(status); +} + +GGML_CALL static void ggml_backend_rpc_buffer_get_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * tensor, void * data, size_t offset, size_t size) { + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + // input serialization format: | rpc_tensor | offset (8 bytes) | size (8 bytes) | + int input_size = sizeof(rpc_tensor) + 2*sizeof(uint64_t); + std::vector input(input_size, 0); + rpc_tensor rpc_tensor = serialize_tensor(tensor); + memcpy(input.data(), &rpc_tensor, sizeof(rpc_tensor)); + memcpy(input.data() + sizeof(rpc_tensor), &offset, sizeof(offset)); + memcpy(input.data() + sizeof(rpc_tensor) + sizeof(offset), &size, sizeof(size)); + std::vector output; + bool status = send_rpc_cmd(ctx->sock, GET_TENSOR, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == size); + // output serialization format: | data (size bytes) | + memcpy(data, output.data(), size); +} + +GGML_CALL static bool ggml_backend_rpc_buffer_cpy_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * src, ggml_tensor * dst) { + // check if src and dst are on the same server + ggml_backend_buffer_t src_buffer = src->buffer; + ggml_backend_rpc_buffer_context * src_ctx = (ggml_backend_rpc_buffer_context *)src_buffer->context; + ggml_backend_buffer_t dst_buffer = dst->buffer; + ggml_backend_rpc_buffer_context * dst_ctx = (ggml_backend_rpc_buffer_context *)dst_buffer->context; + if (src_ctx->sock != dst_ctx->sock) { + return false; + } + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + // input serialization format: | rpc_tensor src | rpc_tensor dst | + int input_size = 2*sizeof(rpc_tensor); + std::vector input(input_size, 0); + rpc_tensor rpc_src = serialize_tensor(src); + rpc_tensor rpc_dst = serialize_tensor(dst); + memcpy(input.data(), &rpc_src, sizeof(rpc_src)); + memcpy(input.data() + sizeof(rpc_src), &rpc_dst, sizeof(rpc_dst)); + std::vector output; + bool status = send_rpc_cmd(ctx->sock, COPY_TENSOR, input, output); + GGML_ASSERT(status); + // output serialization format: | result (1 byte) | + GGML_ASSERT(output.size() == 1); + return output[0]; +} + +GGML_CALL static void ggml_backend_rpc_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { + ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context; + // serialization format: | bufptr (8 bytes) | value (1 byte) | + int input_size = sizeof(uint64_t) + sizeof(uint8_t); + std::vector input(input_size, 0); + memcpy(input.data(), &ctx->remote_ptr, sizeof(ctx->remote_ptr)); + memcpy(input.data() + sizeof(ctx->remote_ptr), &value, sizeof(value)); + std::vector output; + bool status = send_rpc_cmd(ctx->sock, BUFFER_CLEAR, input, output); + GGML_ASSERT(status); +} + +static ggml_backend_buffer_i ggml_backend_rpc_buffer_interface = { + /* .get_name = */ ggml_backend_rpc_buffer_get_name, + /* .free_buffer = */ ggml_backend_rpc_buffer_free_buffer, + /* .get_base = */ ggml_backend_rpc_buffer_get_base, + /* .init_tensor = */ ggml_backend_rpc_buffer_init_tensor, + /* .set_tensor = */ ggml_backend_rpc_buffer_set_tensor, + /* .get_tensor = */ ggml_backend_rpc_buffer_get_tensor, + /* .cpy_tensor = */ ggml_backend_rpc_buffer_cpy_tensor, + /* .clear = */ ggml_backend_rpc_buffer_clear, + /* .reset = */ NULL, +}; + +GGML_CALL static const char * ggml_backend_rpc_buffer_type_name(ggml_backend_buffer_type_t buft) { + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context; + return buft_ctx->name.c_str(); +} + +GGML_CALL static ggml_backend_buffer_t ggml_backend_rpc_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context; + // input serialization format: | size (8 bytes) | + int input_size = sizeof(uint64_t); + std::vector input(input_size, 0); + memcpy(input.data(), &size, sizeof(size)); + std::vector output; + bool status = send_rpc_cmd(buft_ctx->sock, ALLOC_BUFFER, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == 2*sizeof(uint64_t)); + // output serialization format: | remote_ptr (8 bytes) | remote_size (8 bytes) | + uint64_t remote_ptr; + memcpy(&remote_ptr, output.data(), sizeof(remote_ptr)); + size_t remote_size; + memcpy(&remote_size, output.data() + sizeof(uint64_t), sizeof(remote_size)); + + ggml_backend_buffer_t buffer = ggml_backend_buffer_init(buft, + ggml_backend_rpc_buffer_interface, + new ggml_backend_rpc_buffer_context{buft_ctx->sock, {}, remote_ptr, "RPC"}, + remote_size); + + return buffer; +} + +static size_t get_alignment(const std::shared_ptr & sock) { + // input serialization format: | 0 bytes | + std::vector input; + std::vector output; + bool status = send_rpc_cmd(sock, GET_ALIGNMENT, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == sizeof(uint64_t)); + // output serialization format: | alignment (8 bytes) | + uint64_t alignment; + memcpy(&alignment, output.data(), sizeof(alignment)); + return alignment; +} + +GGML_CALL static size_t ggml_backend_rpc_buffer_type_get_alignment(ggml_backend_buffer_type_t buft) { + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context; + return buft_ctx->alignment; +} + +static size_t get_max_size(const std::shared_ptr & sock) { + // input serialization format: | 0 bytes | + std::vector input; + std::vector output; + bool status = send_rpc_cmd(sock, GET_MAX_SIZE, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == sizeof(uint64_t)); + // output serialization format: | max_size (8 bytes) | + uint64_t max_size; + memcpy(&max_size, output.data(), sizeof(max_size)); + return max_size; +} + +GGML_CALL static size_t ggml_backend_rpc_get_max_size(ggml_backend_buffer_type_t buft) { + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context; + return buft_ctx->max_size; +} + +GGML_CALL static size_t ggml_backend_rpc_buffer_type_get_alloc_size(ggml_backend_buffer_type_t buft, const ggml_tensor * tensor) { + UNUSED(buft); + return ggml_nbytes(tensor); +} + +GGML_CALL static bool ggml_backend_rpc_buffer_type_supports_backend(ggml_backend_buffer_type_t buft, ggml_backend_t backend) { + if (!ggml_backend_is_rpc(backend)) { + return false; + } + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context; + ggml_backend_rpc_context * rpc_ctx = (ggml_backend_rpc_context *)backend->context; + return buft_ctx->sock == rpc_ctx->sock; +} + +static ggml_backend_buffer_type_i ggml_backend_rpc_buffer_type_interface = { + /* .get_name = */ ggml_backend_rpc_buffer_type_name, + /* .alloc_buffer = */ ggml_backend_rpc_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_rpc_buffer_type_get_alignment, + /* .get_max_size = */ ggml_backend_rpc_get_max_size, + /* .get_alloc_size = */ ggml_backend_rpc_buffer_type_get_alloc_size, + /* .supports_backend = */ ggml_backend_rpc_buffer_type_supports_backend, + /* .is_host = */ NULL, +}; + + +GGML_CALL static const char * ggml_backend_rpc_name(ggml_backend_t backend) { + ggml_backend_rpc_context * rpc_ctx = (ggml_backend_rpc_context *)backend->context; + + return rpc_ctx->name.c_str(); +} + +GGML_CALL static void ggml_backend_rpc_free(ggml_backend_t backend) { + ggml_backend_rpc_context * rpc_ctx = (ggml_backend_rpc_context *)backend->context; + ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)rpc_ctx->buft->context; + delete buft_ctx; + delete rpc_ctx->buft; + delete rpc_ctx; + delete backend; +} + +GGML_CALL static ggml_backend_buffer_type_t ggml_backend_rpc_get_default_buffer_type(ggml_backend_t backend) { + ggml_backend_rpc_context * ctx = (ggml_backend_rpc_context *)backend->context; + return ctx->buft; +} + +GGML_CALL static void ggml_backend_rpc_synchronize(ggml_backend_t backend) { + UNUSED(backend); + // this is no-op because we don't have any async operations +} + +static void add_tensor(ggml_tensor * tensor, std::vector & tensors, std::unordered_set & visited) { + if (tensor == nullptr) { + return; + } + if (visited.find(tensor) != visited.end()) { + return; + } + visited.insert(tensor); + for (int i = 0; i < GGML_MAX_SRC; i++) { + add_tensor(tensor->src[i], tensors, visited); + } + add_tensor(tensor->view_src, tensors, visited); + tensors.push_back(serialize_tensor(tensor)); +} + +static void serialize_graph(const ggml_cgraph * cgraph, std::vector & output) { + uint32_t n_nodes = cgraph->n_nodes; + std::vector tensors; + std::unordered_set visited; + for (uint32_t i = 0; i < n_nodes; i++) { + add_tensor(cgraph->nodes[i], tensors, visited); + } + // serialization format: + // | n_nodes (4 bytes) | nodes (n_nodes * sizeof(uint64_t) | n_tensors (4 bytes) | tensors (n_tensors * sizeof(rpc_tensor)) | + uint32_t n_tensors = tensors.size(); + int output_size = sizeof(uint32_t) + n_nodes * sizeof(uint64_t) + sizeof(uint32_t) + n_tensors * sizeof(rpc_tensor); + output.resize(output_size, 0); + memcpy(output.data(), &n_nodes, sizeof(n_nodes)); + uint64_t * out_nodes = (uint64_t *)(output.data() + sizeof(n_nodes)); + for (uint32_t i = 0; i < n_nodes; i++) { + out_nodes[i] = reinterpret_cast(cgraph->nodes[i]); + } + uint32_t * out_ntensors = (uint32_t *)(output.data() + sizeof(n_nodes) + n_nodes * sizeof(uint64_t)); + *out_ntensors = n_tensors; + rpc_tensor * out_tensors = (rpc_tensor *)(output.data() + sizeof(n_nodes) + n_nodes * sizeof(uint64_t) + sizeof(uint32_t)); + memcpy(out_tensors, tensors.data(), n_tensors * sizeof(rpc_tensor)); +} + +GGML_CALL static enum ggml_status ggml_backend_rpc_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) { + ggml_backend_rpc_context * rpc_ctx = (ggml_backend_rpc_context *)backend->context; + std::vector input; + serialize_graph(cgraph, input); + std::vector output; + bool status = send_rpc_cmd(rpc_ctx->sock, GRAPH_COMPUTE, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == 1); + return (enum ggml_status)output[0]; +} + +GGML_CALL static bool ggml_backend_rpc_supports_op(ggml_backend_t backend, const ggml_tensor * op) { + UNUSED(backend); + UNUSED(op); + GGML_ASSERT(false && "not implemented"); + return false; +} + +static ggml_backend_i ggml_backend_rpc_interface = { + /* .get_name = */ ggml_backend_rpc_name, + /* .free = */ ggml_backend_rpc_free, + /* .get_default_buffer_type = */ ggml_backend_rpc_get_default_buffer_type, + /* .set_tensor_async = */ NULL, + /* .get_tensor_async = */ NULL, + /* .cpy_tensor_async = */ NULL, + /* .synchronize = */ ggml_backend_rpc_synchronize, + /* .graph_plan_create = */ NULL, + /* .graph_plan_free = */ NULL, + /* .graph_plan_compute = */ NULL, + /* .graph_compute = */ ggml_backend_rpc_graph_compute, + /* .supports_op = */ ggml_backend_rpc_supports_op, + /* .offload_op = */ NULL, + /* .event_new = */ NULL, + /* .event_free = */ NULL, + /* .event_record = */ NULL, + /* .event_wait = */ NULL, + /* .event_synchronize = */ NULL, +}; + +static std::unordered_map instances; + +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint) { + ggml_backend_t backend = ggml_backend_rpc_init(endpoint); + return backend != nullptr ? ggml_backend_rpc_get_default_buffer_type(backend) : nullptr; +} + +GGML_CALL ggml_backend_t ggml_backend_rpc_init(const char * endpoint) { + std::string endpoint_str(endpoint); + if (instances.find(endpoint_str) != instances.end()) { + return instances[endpoint_str]; + } +#ifdef _WIN32 + { + WSADATA wsaData; + int res = WSAStartup(MAKEWORD(2, 2), &wsaData); + if (res != 0) { + return nullptr; + } + } +#endif + GGML_PRINT_DEBUG("Connecting to %s\n", endpoint); + std::string host; + int port; + if (!parse_endpoint(endpoint, host, port)) { + return nullptr; + } + auto sock = socket_connect(host.c_str(), port); + if (sock == nullptr) { + return nullptr; + } + size_t alignment = get_alignment(sock); + size_t max_size = get_max_size(sock); + ggml_backend_rpc_buffer_type_context * buft_ctx = new ggml_backend_rpc_buffer_type_context { + /* .sock = */ sock, + /* .name = */ "RPC" + std::to_string(sock->fd), + /* .alignment = */ alignment, + /* .max_size = */ max_size + }; + + ggml_backend_buffer_type_t buft = new ggml_backend_buffer_type { + /* .iface = */ ggml_backend_rpc_buffer_type_interface, + /* .context = */ buft_ctx + }; + + ggml_backend_rpc_context * ctx = new ggml_backend_rpc_context { + /* .endpoint = */ endpoint, + /* .name = */ "RPC" + std::to_string(sock->fd), + /* .sock = */ sock, + /* .buft = */ buft + }; + + instances[endpoint] = new ggml_backend { + /* .guid = */ ggml_backend_rpc_guid(), + /* .interface = */ ggml_backend_rpc_interface, + /* .context = */ ctx + }; + + return instances[endpoint]; +} + +GGML_API GGML_CALL bool ggml_backend_is_rpc(ggml_backend_t backend) { + return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_rpc_guid()); +} + +static void get_device_memory(const std::shared_ptr & sock, size_t * free, size_t * total) { + // input serialization format: | 0 bytes | + std::vector input; + std::vector output; + bool status = send_rpc_cmd(sock, GET_DEVICE_MEMORY, input, output); + GGML_ASSERT(status); + GGML_ASSERT(output.size() == 2*sizeof(uint64_t)); + // output serialization format: | free (8 bytes) | total (8 bytes) | + uint64_t free_mem; + memcpy(&free_mem, output.data(), sizeof(free_mem)); + uint64_t total_mem; + memcpy(&total_mem, output.data() + sizeof(uint64_t), sizeof(total_mem)); + *free = free_mem; + *total = total_mem; +} + +GGML_API GGML_CALL void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total) { + ggml_backend_t backend = ggml_backend_rpc_init(endpoint); + if (backend == nullptr) { + *free = 0; + *total = 0; + return; + } + ggml_backend_rpc_context * ctx = (ggml_backend_rpc_context *)backend->context; + get_device_memory(ctx->sock, free, total); +} + +// RPC server-side implementation + +static void rpc_alloc_buffer(ggml_backend_t backend, const std::vector & input, std::vector & output) { + // input serialization format: | size (8 bytes) | + uint64_t size; + memcpy(&size, input.data(), sizeof(size)); + ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend); + ggml_backend_buffer_t buffer = ggml_backend_buft_alloc_buffer(buft, size); + uint64_t remote_ptr = reinterpret_cast(buffer); + uint64_t remote_size = buffer->size; + GGML_PRINT_DEBUG("[%s] size: %" PRIu64 " -> remote_ptr: %" PRIx64 ", remote_size: %" PRIu64 "\n", __func__, size, remote_ptr, remote_size); + // output serialization format: | remote_ptr (8 bytes) | remote_size (8 bytes) | + output.resize(2*sizeof(uint64_t), 0); + memcpy(output.data(), &remote_ptr, sizeof(remote_ptr)); + memcpy(output.data() + sizeof(uint64_t), &remote_size, sizeof(remote_size)); +} + +static void rpc_get_alignment(ggml_backend_t backend, std::vector & output) { + ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend); + size_t alignment = ggml_backend_buft_get_alignment(buft); + GGML_PRINT_DEBUG("[%s] alignment: %lu\n", __func__, alignment); + // output serialization format: | alignment (8 bytes) | + output.resize(sizeof(uint64_t), 0); + memcpy(output.data(), &alignment, sizeof(alignment)); +} + +static void rpc_get_max_size(ggml_backend_t backend, std::vector & output) { + ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend); + size_t max_size = ggml_backend_buft_get_max_size(buft); + GGML_PRINT_DEBUG("[%s] max_size: %lu\n", __func__, max_size); + // output serialization format: | max_size (8 bytes) | + output.resize(sizeof(uint64_t), 0); + memcpy(output.data(), &max_size, sizeof(max_size)); +} + +static void rpc_buffer_get_base(const std::vector & input, std::vector & output) { + // input serialization format: | remote_ptr (8 bytes) | + uint64_t remote_ptr; + memcpy(&remote_ptr, input.data(), sizeof(remote_ptr)); + GGML_PRINT_DEBUG("[%s] remote_ptr: %" PRIx64 "\n", __func__, remote_ptr); + ggml_backend_buffer_t buffer = reinterpret_cast(remote_ptr); + void * base = ggml_backend_buffer_get_base(buffer); + // output serialization format: | base_ptr (8 bytes) | + uint64_t base_ptr = reinterpret_cast(base); + output.resize(sizeof(uint64_t), 0); + memcpy(output.data(), &base_ptr, sizeof(base_ptr)); +} + +static void rpc_free_buffer(const std::vector & input) { + // input serialization format: | remote_ptr (8 bytes) | + uint64_t remote_ptr; + memcpy(&remote_ptr, input.data(), sizeof(remote_ptr)); + GGML_PRINT_DEBUG("[%s] remote_ptr: %" PRIx64 "\n", __func__, remote_ptr); + ggml_backend_buffer_t buffer = reinterpret_cast(remote_ptr); + ggml_backend_buffer_free(buffer); +} + +static void rpc_buffer_clear(const std::vector & input) { + // input serialization format: | remote_ptr (8 bytes) | value (1 byte) | + uint64_t remote_ptr; + memcpy(&remote_ptr, input.data(), sizeof(remote_ptr)); + uint8_t value; + memcpy(&value, input.data() + sizeof(uint64_t), sizeof(value)); + GGML_PRINT_DEBUG("[%s] remote_ptr: %" PRIx64 ", value: %u\n", __func__, remote_ptr, value); + ggml_backend_buffer_t buffer = reinterpret_cast(remote_ptr); + ggml_backend_buffer_clear(buffer, value); +} + +static void rpc_set_tensor(const std::vector & input) { + // serialization format: | rpc_tensor | offset (8 bytes) | data (size bytes) | + const rpc_tensor * in_tensor = (const rpc_tensor *)input.data(); + uint64_t offset; + memcpy(&offset, input.data() + sizeof(rpc_tensor), sizeof(offset)); + size_t size = input.size() - sizeof(rpc_tensor) - sizeof(offset); + + struct ggml_init_params params { + /*.mem_size =*/ ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx = ggml_init(params); + ggml_tensor * tensor = deserialize_tensor(ctx, in_tensor); + GGML_PRINT_DEBUG("[%s] buffer: %p, data: %p, offset: %" PRIu64 ", size: %zu\n", __func__, (void*)tensor->buffer, tensor->data, offset, size); + const void * data = input.data() + sizeof(rpc_tensor) + sizeof(offset); + ggml_backend_tensor_set(tensor, data, offset, size); + ggml_free(ctx); +} + +static void rpc_get_tensor(const std::vector & input, std::vector & output) { + // serialization format: | rpc_tensor | offset (8 bytes) | size (8 bytes) | + const rpc_tensor * in_tensor = (const rpc_tensor *)input.data(); + uint64_t offset; + memcpy(&offset, input.data() + sizeof(rpc_tensor), sizeof(offset)); + uint64_t size; + memcpy(&size, input.data() + sizeof(rpc_tensor) + sizeof(offset), sizeof(size)); + + struct ggml_init_params params { + /*.mem_size =*/ ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx = ggml_init(params); + ggml_tensor * tensor = deserialize_tensor(ctx, in_tensor); + GGML_PRINT_DEBUG("[%s] buffer: %p, data: %p, offset: %" PRIu64 ", size: %" PRIu64 "\n", __func__, (void*)tensor->buffer, tensor->data, offset, size); + // output serialization format: | data (size bytes) | + output.resize(size, 0); + ggml_backend_tensor_get(tensor, output.data(), offset, size); + ggml_free(ctx); +} + +static void rpc_copy_tensor(const std::vector & input, std::vector & output) { + // serialization format: | rpc_tensor src | rpc_tensor dst | + const rpc_tensor * rpc_src = (const rpc_tensor *)input.data(); + const rpc_tensor * rpc_dst = (const rpc_tensor *)(input.data() + sizeof(rpc_src)); + + struct ggml_init_params params { + /*.mem_size =*/ 2*ggml_tensor_overhead(), + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx = ggml_init(params); + ggml_tensor * src = deserialize_tensor(ctx, rpc_src); + ggml_tensor * dst = deserialize_tensor(ctx, rpc_dst); + GGML_PRINT_DEBUG("[%s] src->buffer: %p, dst->buffer: %p\n", __func__, (void*)src->buffer, (void*)dst->buffer); + bool result = ggml_backend_buffer_copy_tensor(src, dst); + // output serialization format: | result (1 byte) | + output.resize(1, 0); + output[0] = result; + ggml_free(ctx); +} + +static struct ggml_tensor * create_node(uint64_t id, + struct ggml_context * ctx, + const std::unordered_map & tensor_ptrs, + std::unordered_map & tensor_map) { + if (id == 0) { + return nullptr; + } + if (tensor_map.find(id) != tensor_map.end()) { + return tensor_map[id]; + } + const rpc_tensor * tensor = tensor_ptrs.at(id); + struct ggml_tensor * result = deserialize_tensor(ctx, tensor); + tensor_map[id] = result; + for (int i = 0; i < GGML_MAX_SRC; i++) { + result->src[i] = create_node(tensor->src[i], ctx, tensor_ptrs, tensor_map); + } + result->view_src = create_node(tensor->view_src, ctx, tensor_ptrs, tensor_map); + result->view_offs = tensor->view_offs; + return result; +} + +static void rpc_graph_compute(ggml_backend_t backend, const std::vector & input, std::vector & output) { + // serialization format: + // | n_nodes (4 bytes) | nodes (n_nodes * sizeof(uint64_t) | n_tensors (4 bytes) | tensors (n_tensors * sizeof(rpc_tensor)) | + uint32_t n_nodes; + memcpy(&n_nodes, input.data(), sizeof(n_nodes)); + const uint64_t * nodes = (const uint64_t *)(input.data() + sizeof(n_nodes)); + uint32_t n_tensors; + memcpy(&n_tensors, input.data() + sizeof(n_nodes) + n_nodes*sizeof(uint64_t), sizeof(n_tensors)); + const rpc_tensor * tensors = (const rpc_tensor *)(input.data() + sizeof(n_nodes) + n_nodes*sizeof(uint64_t) + sizeof(n_tensors)); + GGML_PRINT_DEBUG("[%s] n_nodes: %u, n_tensors: %u\n", __func__, n_nodes, n_tensors); + + static size_t buf_size = ggml_tensor_overhead()*(n_nodes + n_tensors) + ggml_graph_overhead_custom(n_nodes, false); + struct ggml_init_params params = { + /*.mem_size =*/ buf_size, + /*.mem_buffer =*/ NULL, + /*.no_alloc =*/ true, + }; + struct ggml_context * ctx = ggml_init(params); + struct ggml_cgraph * graph = ggml_new_graph_custom(ctx, n_nodes, false); + graph->n_nodes = n_nodes; + std::unordered_map tensor_ptrs; + for (uint32_t i = 0; i < n_tensors; i++) { + tensor_ptrs[tensors[i].id] = &tensors[i]; + } + std::unordered_map tensor_map; + for (uint32_t i = 0; i < n_nodes; i++) { + graph->nodes[i] = create_node(nodes[i], ctx, tensor_ptrs, tensor_map); + } + ggml_status status = ggml_backend_graph_compute(backend, graph); + // output serialization format: | status (1 byte) | + output.resize(1, 0); + output[0] = status; + ggml_free(ctx); +} + +static void rpc_serve_client(ggml_backend_t backend, sockfd_t sockfd, size_t free_mem, size_t total_mem) { + while (true) { + uint8_t cmd; + if (!recv_data(sockfd, &cmd, 1)) { + break; + } + std::vector input; + std::vector output; + uint64_t input_size; + if (!recv_data(sockfd, &input_size, sizeof(input_size))) { + break; + } + input.resize(input_size); + if (!recv_data(sockfd, input.data(), input_size)) { + break; + } + switch (cmd) { + case ALLOC_BUFFER: { + rpc_alloc_buffer(backend, input, output); + break; + } + case GET_ALIGNMENT: { + rpc_get_alignment(backend, output); + break; + } + case GET_MAX_SIZE: { + rpc_get_max_size(backend, output); + break; + } + case BUFFER_GET_BASE: { + rpc_buffer_get_base(input, output); + break; + } + case FREE_BUFFER: { + rpc_free_buffer(input); + break; + } + case BUFFER_CLEAR: { + rpc_buffer_clear(input); + break; + } + case SET_TENSOR: { + rpc_set_tensor(input); + break; + } + case GET_TENSOR: { + rpc_get_tensor(input, output); + break; + } + case COPY_TENSOR: { + rpc_copy_tensor(input, output); + break; + } + case GRAPH_COMPUTE: { + rpc_graph_compute(backend, input, output); + break; + } + case GET_DEVICE_MEMORY: { + // output serialization format: | free (8 bytes) | total (8 bytes) | + output.resize(2*sizeof(uint64_t), 0); + memcpy(output.data(), &free_mem, sizeof(free_mem)); + memcpy(output.data() + sizeof(uint64_t), &total_mem, sizeof(total_mem)); + break; + } + default: { + fprintf(stderr, "Unknown command: %d\n", cmd); + return; + } + } + uint64_t output_size = output.size(); + if (!send_data(sockfd, &output_size, sizeof(output_size))) { + break; + } + if (!send_data(sockfd, output.data(), output_size)) { + break; + } + } +} + +void start_rpc_server(ggml_backend_t backend, const char * endpoint, size_t free_mem, size_t total_mem) { + std::string host; + int port; + if (!parse_endpoint(endpoint, host, port)) { + return; + } +#ifdef _WIN32 + { + WSADATA wsaData; + int res = WSAStartup(MAKEWORD(2, 2), &wsaData); + if (res != 0) { + fprintf(stderr, "WSAStartup failed: %d\n", res); + return; + } + } +#endif + auto server_socket = create_server_socket(host.c_str(), port); + if (server_socket == nullptr) { + fprintf(stderr, "Failed to create server socket\n"); + return; + } + while (true) { + auto client_socket = socket_accept(server_socket->fd); + if (client_socket == nullptr) { + fprintf(stderr, "Failed to accept client connection\n"); + return; + } + printf("Accepted client connection, free_mem=%zu, total_mem=%zu\n", free_mem, total_mem); + rpc_serve_client(backend, client_socket->fd, free_mem, total_mem); + printf("Client connection closed\n"); + } +#ifdef _WIN32 + WSACleanup(); +#endif +} diff --git a/ggml-rpc.h b/ggml-rpc.h new file mode 100644 index 00000000000..aa144832a6e --- /dev/null +++ b/ggml-rpc.h @@ -0,0 +1,24 @@ +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#ifdef __cplusplus +extern "C" { +#endif + +#define GGML_RPC_MAX_SERVERS 16 + +// backend API +GGML_API GGML_CALL ggml_backend_t ggml_backend_rpc_init(const char * endpoint); +GGML_API GGML_CALL bool ggml_backend_is_rpc(ggml_backend_t backend); + +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint); + +GGML_API GGML_CALL void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total); + +GGML_API GGML_CALL void start_rpc_server(ggml_backend_t backend, const char * endpoint, size_t free_mem, size_t total_mem); + +#ifdef __cplusplus +} +#endif From 1056ad762cdc15141e4eba3db5bd6e83eaa4b28f Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Tue, 14 May 2024 19:09:30 +0300 Subject: [PATCH 05/17] metal : support FA without mask + add asserts (llama/7278) * ggml : fa without mask + add asserts ggml-ci * metal : support non-contiguous KV ggml-ci --- ggml-metal.m | 69 ++++++++++++++++++++++++++---------------------- ggml-metal.metal | 53 ++++++++++++++----------------------- ggml.c | 10 +++++++ ggml.h | 3 ++- 4 files changed, 70 insertions(+), 65 deletions(-) diff --git a/ggml-metal.m b/ggml-metal.m index bfa352c3a9a..390a1cd7890 100644 --- a/ggml-metal.m +++ b/ggml-metal.m @@ -2512,13 +2512,14 @@ static enum ggml_status ggml_metal_graph_compute( } break; case GGML_OP_FLASH_ATTN_EXT: { - GGML_ASSERT(ne00 % 4 == 0); + GGML_ASSERT(ne00 % 4 == 0); + GGML_ASSERT(ne11 % 32 == 0); + GGML_ASSERT(src0->type == GGML_TYPE_F32); - struct ggml_tensor * src3 = gf->nodes[i]->src[3]; + GGML_ASSERT(ggml_are_same_shape (src1, src2)); - GGML_ASSERT(ggml_are_same_shape(src1, src2)); - GGML_ASSERT(src3); + struct ggml_tensor * src3 = gf->nodes[i]->src[3]; size_t offs_src3 = 0; @@ -2528,6 +2529,11 @@ static enum ggml_status ggml_metal_graph_compute( GGML_ASSERT(!src3 || src3->ne[1] >= GGML_PAD(src0->ne[1], 8) && "the Flash-Attention Metal kernel requires the mask to be padded to 8 and at least n_queries big"); + const uint64_t nb20 = src2 ? src2->nb[0] : 0; GGML_UNUSED(nb20); + const uint64_t nb21 = src2 ? src2->nb[1] : 0; + const uint64_t nb22 = src2 ? src2->nb[2] : 0; + const uint64_t nb23 = src2 ? src2->nb[3] : 0; + const int64_t ne30 = src3 ? src3->ne[0] : 0; GGML_UNUSED(ne30); //const int64_t ne31 = src3 ? src3->ne[1] : 0; const int64_t ne32 = src3 ? src3->ne[2] : 0; GGML_UNUSED(ne32); @@ -2590,34 +2596,35 @@ static enum ggml_status ggml_metal_graph_compute( [encoder setBuffer:id_src0 offset:offs_src0 atIndex:0]; [encoder setBuffer:id_src1 offset:offs_src1 atIndex:1]; [encoder setBuffer:id_src2 offset:offs_src2 atIndex:2]; - [encoder setBuffer:id_src3 offset:offs_src3 atIndex:3]; + if (id_src3) { + [encoder setBuffer:id_src3 offset:offs_src3 atIndex:3]; + } else { + [encoder setBuffer:id_src0 offset:offs_src0 atIndex:3]; + } [encoder setBuffer:id_dst offset:offs_dst atIndex:4]; - [encoder setBytes:&ne00 length:sizeof( int64_t) atIndex:5]; - [encoder setBytes:&ne01 length:sizeof( int64_t) atIndex:6]; - [encoder setBytes:&ne02 length:sizeof( int64_t) atIndex:7]; - [encoder setBytes:&ne03 length:sizeof( int64_t) atIndex:8]; - [encoder setBytes:&nb00 length:sizeof(uint64_t) atIndex:9]; - [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:10]; - [encoder setBytes:&nb02 length:sizeof(uint64_t) atIndex:11]; - [encoder setBytes:&nb03 length:sizeof(uint64_t) atIndex:12]; - [encoder setBytes:&ne10 length:sizeof( int64_t) atIndex:13]; - [encoder setBytes:&ne11 length:sizeof( int64_t) atIndex:14]; - [encoder setBytes:&ne12 length:sizeof( int64_t) atIndex:15]; - [encoder setBytes:&ne13 length:sizeof( int64_t) atIndex:16]; - [encoder setBytes:&nb10 length:sizeof(uint64_t) atIndex:17]; - [encoder setBytes:&nb11 length:sizeof(uint64_t) atIndex:18]; - [encoder setBytes:&nb12 length:sizeof(uint64_t) atIndex:19]; - [encoder setBytes:&nb13 length:sizeof(uint64_t) atIndex:20]; - [encoder setBytes:&nb31 length:sizeof(uint64_t) atIndex:21]; - [encoder setBytes:&ne0 length:sizeof( int64_t) atIndex:22]; - [encoder setBytes:&ne1 length:sizeof( int64_t) atIndex:23]; - [encoder setBytes:&ne2 length:sizeof( int64_t) atIndex:24]; - [encoder setBytes:&ne3 length:sizeof( int64_t) atIndex:25]; - [encoder setBytes:&scale length:sizeof( float) atIndex:26]; - [encoder setBytes:&max_bias length:sizeof( float) atIndex:27]; - [encoder setBytes:&m0 length:sizeof(m0) atIndex:28]; - [encoder setBytes:&m1 length:sizeof(m1) atIndex:29]; - [encoder setBytes:&n_head_log2 length:sizeof(n_head_log2) atIndex:30]; + [encoder setBytes:&ne01 length:sizeof( int64_t) atIndex:5]; + [encoder setBytes:&ne02 length:sizeof( int64_t) atIndex:6]; + [encoder setBytes:&ne03 length:sizeof( int64_t) atIndex:7]; + [encoder setBytes:&nb01 length:sizeof(uint64_t) atIndex:8]; + [encoder setBytes:&nb02 length:sizeof(uint64_t) atIndex:9]; + [encoder setBytes:&nb03 length:sizeof(uint64_t) atIndex:10]; + [encoder setBytes:&ne11 length:sizeof( int64_t) atIndex:11]; + [encoder setBytes:&ne12 length:sizeof( int64_t) atIndex:12]; + [encoder setBytes:&ne13 length:sizeof( int64_t) atIndex:13]; + [encoder setBytes:&nb11 length:sizeof(uint64_t) atIndex:14]; + [encoder setBytes:&nb12 length:sizeof(uint64_t) atIndex:15]; + [encoder setBytes:&nb13 length:sizeof(uint64_t) atIndex:16]; + [encoder setBytes:&nb21 length:sizeof(uint64_t) atIndex:17]; + [encoder setBytes:&nb22 length:sizeof(uint64_t) atIndex:18]; + [encoder setBytes:&nb23 length:sizeof(uint64_t) atIndex:19]; + [encoder setBytes:&nb31 length:sizeof(uint64_t) atIndex:20]; + [encoder setBytes:&ne1 length:sizeof( int64_t) atIndex:21]; + [encoder setBytes:&ne2 length:sizeof( int64_t) atIndex:22]; + [encoder setBytes:&scale length:sizeof( float) atIndex:23]; + [encoder setBytes:&max_bias length:sizeof( float) atIndex:24]; + [encoder setBytes:&m0 length:sizeof(m0) atIndex:25]; + [encoder setBytes:&m1 length:sizeof(m1) atIndex:26]; + [encoder setBytes:&n_head_log2 length:sizeof(n_head_log2) atIndex:27]; if (!use_vec_kernel) { // half8x8 kernel diff --git a/ggml-metal.metal b/ggml-metal.metal index 7af4e8f9342..57fdf564e17 100644 --- a/ggml-metal.metal +++ b/ggml-metal.metal @@ -2049,27 +2049,24 @@ typedef void (flash_attn_ext_f16_t)( device const char * v, device const char * mask, device float * dst, - constant int64_t & ne00, constant int64_t & ne01, constant int64_t & ne02, constant int64_t & ne03, - constant uint64_t & nb00, constant uint64_t & nb01, constant uint64_t & nb02, constant uint64_t & nb03, - constant int64_t & ne10, constant int64_t & ne11, constant int64_t & ne12, constant int64_t & ne13, - constant uint64_t & nb10, constant uint64_t & nb11, constant uint64_t & nb12, constant uint64_t & nb13, + constant uint64_t & nb21, + constant uint64_t & nb22, + constant uint64_t & nb23, constant uint64_t & nb31, - constant int64_t & ne0, constant int64_t & ne1, constant int64_t & ne2, - constant int64_t & ne3, constant float & scale, constant float & max_bias, constant float & m0, @@ -2090,27 +2087,24 @@ kernel void kernel_flash_attn_ext_f16( device const char * v, device const char * mask, device float * dst, - constant int64_t & ne00, constant int64_t & ne01, constant int64_t & ne02, constant int64_t & ne03, - constant uint64_t & nb00, constant uint64_t & nb01, constant uint64_t & nb02, constant uint64_t & nb03, - constant int64_t & ne10, constant int64_t & ne11, constant int64_t & ne12, constant int64_t & ne13, - constant uint64_t & nb10, constant uint64_t & nb11, constant uint64_t & nb12, constant uint64_t & nb13, + constant uint64_t & nb21, + constant uint64_t & nb22, + constant uint64_t & nb23, constant uint64_t & nb31, - constant int64_t & ne0, constant int64_t & ne1, constant int64_t & ne2, - constant int64_t & ne3, constant float & scale, constant float & max_bias, constant float & m0, @@ -2180,10 +2174,6 @@ kernel void kernel_flash_attn_ext_f16( const short ne22 = ne12; const short ne23 = ne13; - const uint nb21 = nb11; - const uint nb22 = nb12; - const uint nb23 = nb13; - // broadcast const short rk2 = ne02/ne12; const short rk3 = ne03/ne13; @@ -2247,11 +2237,16 @@ kernel void kernel_flash_attn_ext_f16( simdgroup_multiply_accumulate(mqk, mq[i], mk, mqk); } - // mqk = mqk*scale + mask*slope - simdgroup_half8x8 mm; - simdgroup_load(mm, mp + ic + 8*cc, nb31/sizeof(half), 0, false); - simdgroup_multiply(mm, mslope, mm); - simdgroup_multiply_accumulate(mqk, mqk, mscale, mm); + if (mask != q) { + // mqk = mqk*scale + mask*slope + simdgroup_half8x8 mm; + simdgroup_load(mm, mp + ic + 8*cc, nb31/sizeof(half), 0, false); + simdgroup_multiply(mm, mslope, mm); + simdgroup_multiply_accumulate(mqk, mqk, mscale, mm); + } else { + // mqk = mqk*scale + simdgroup_multiply(mqk, mscale, mqk); + } simdgroup_store(mqk, ss + 8*cc, TF, 0, false); } @@ -2425,27 +2420,24 @@ kernel void kernel_flash_attn_ext_vec_f16( device const char * v, device const char * mask, device float * dst, - constant int64_t & ne00, constant int64_t & ne01, constant int64_t & ne02, constant int64_t & ne03, - constant uint64_t & nb00, constant uint64_t & nb01, constant uint64_t & nb02, constant uint64_t & nb03, - constant int64_t & ne10, constant int64_t & ne11, constant int64_t & ne12, constant int64_t & ne13, - constant uint64_t & nb10, constant uint64_t & nb11, constant uint64_t & nb12, constant uint64_t & nb13, + constant uint64_t & nb21, + constant uint64_t & nb22, + constant uint64_t & nb23, constant uint64_t & nb31, - constant int64_t & ne0, constant int64_t & ne1, constant int64_t & ne2, - constant int64_t & ne3, constant float & scale, constant float & max_bias, constant float & m0, @@ -2521,10 +2513,6 @@ kernel void kernel_flash_attn_ext_vec_f16( const short ne22 = ne12; const short ne23 = ne13; - const uint nb21 = nb11; - const uint nb22 = nb12; - const uint nb23 = nb13; - // broadcast const short rk2 = ne02/ne12; const short rk3 = ne03/ne13; @@ -2589,8 +2577,7 @@ kernel void kernel_flash_attn_ext_vec_f16( // mqk = mqk*scale + mask*slope if (tiisg == 0) { - float4 mm = (float4) mp4[ic/4 + cc]; - mqk = mqk*scale + mm*slope; + mqk = mqk*scale + ((mask != q) ? ((float4) mp4[ic/4 + cc])*slope : (float4) 0.0f); ss4[cc] = mqk; } diff --git a/ggml.c b/ggml.c index d443a9b42ce..03b609dddce 100644 --- a/ggml.c +++ b/ggml.c @@ -2824,6 +2824,16 @@ bool ggml_are_same_shape(const struct ggml_tensor * t0, const struct ggml_tensor (t0->ne[3] == t1->ne[3] ); } +bool ggml_are_same_stride(const struct ggml_tensor * t0, const struct ggml_tensor * t1) { + static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function"); + + return + (t0->nb[0] == t1->nb[0] ) && + (t0->nb[1] == t1->nb[1] ) && + (t0->nb[2] == t1->nb[2] ) && + (t0->nb[3] == t1->nb[3] ); +} + // check if t1 can be represented as a repeatition of t0 static inline bool ggml_can_repeat(const struct ggml_tensor * t0, const struct ggml_tensor * t1) { static_assert(GGML_MAX_DIMS == 4, "GGML_MAX_DIMS is not 4 - update this function"); diff --git a/ggml.h b/ggml.h index 3fe95ed5763..25f4f73a8d9 100644 --- a/ggml.h +++ b/ggml.h @@ -766,7 +766,8 @@ extern "C" { GGML_API bool ggml_is_3d (const struct ggml_tensor * tensor); GGML_API int ggml_n_dims (const struct ggml_tensor * tensor); // returns 1 for scalars - GGML_API bool ggml_are_same_shape(const struct ggml_tensor * t0, const struct ggml_tensor * t1); + GGML_API bool ggml_are_same_shape (const struct ggml_tensor * t0, const struct ggml_tensor * t1); + GGML_API bool ggml_are_same_stride(const struct ggml_tensor * t0, const struct ggml_tensor * t1); // use this to compute the memory overhead of a tensor GGML_API size_t ggml_tensor_overhead(void); From f56b8305c4f5760b5612a93305ed57aef082bfa5 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Tue, 14 May 2024 19:16:32 +0300 Subject: [PATCH 06/17] sync : ggml --- scripts/sync-ggml.last | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/sync-ggml.last b/scripts/sync-ggml.last index 0096c0b533a..35eef4660fd 100644 --- a/scripts/sync-ggml.last +++ b/scripts/sync-ggml.last @@ -1 +1 @@ -9149580f5e15fa7510fa3413516fbf517cf2e921 +e87c0557b012350005269c49e1c2b5a8631da59a From 9d5771ae43d7fc7cca9d31dd924b13a29144e476 Mon Sep 17 00:00:00 2001 From: petterreinholdtsen Date: Tue, 14 May 2024 20:32:41 +0200 Subject: [PATCH 07/17] talk-llama : reject runs without required arguments (#2153) * Extended talk-llama example to reject runs without required arguments. Print warning and exit if models are not specified on the command line. * Update examples/talk-llama/talk-llama.cpp * Update examples/talk-llama/talk-llama.cpp --------- Co-authored-by: Georgi Gerganov --- examples/talk-llama/talk-llama.cpp | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/examples/talk-llama/talk-llama.cpp b/examples/talk-llama/talk-llama.cpp index bb8c26d5efd..838d6f56357 100644 --- a/examples/talk-llama/talk-llama.cpp +++ b/examples/talk-llama/talk-llama.cpp @@ -288,6 +288,10 @@ int main(int argc, char ** argv) { cparams.use_gpu = params.use_gpu; struct whisper_context * ctx_wsp = whisper_init_from_file_with_params(params.model_wsp.c_str(), cparams); + if (!ctx_wsp) { + fprintf(stderr, "No whisper.cpp model specified. Please provide using -mw \n"); + return 1; + } // llama init @@ -301,6 +305,10 @@ int main(int argc, char ** argv) { } struct llama_model * model_llama = llama_load_model_from_file(params.model_llama.c_str(), lmparams); + if (!model_llama) { + fprintf(stderr, "No llama.cpp model specified. Please provide using -ml \n"); + return 1; + } llama_context_params lcparams = llama_context_default_params(); From 7094ea5e750266e16c16c7aecac8fc03294ecaa3 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Wed, 15 May 2024 09:38:19 +0300 Subject: [PATCH 08/17] whisper : use flash attention (#2152) * whisper : use flash attention in the encoder * whisper : add kv_pad * whisper : remove extra backend instance (huh?) * whisper : use FA for cross-attention * whisper : use FA for self-attention * whisper : simplify encoder FA * whisper : add flash_attn runtime parameter * scripts : add bench log * scripts : add M1 Pro bench log --- examples/bench/bench.cpp | 17 +- examples/command/command.cpp | 7 +- examples/lsp/lsp.cpp | 8 +- examples/main/main.cpp | 9 +- examples/server/server.cpp | 7 +- examples/stream/stream.cpp | 7 +- examples/talk-llama/talk-llama.cpp | 9 +- examples/talk/talk.cpp | 7 +- examples/wchess/wchess.cmd/wchess.cmd.cpp | 7 +- scripts/bench-all-gg.txt | 298 +++++++++++++++ scripts/bench-all.sh | 25 +- whisper.cpp | 429 ++++++++++++++-------- whisper.h | 1 + 13 files changed, 658 insertions(+), 173 deletions(-) create mode 100644 scripts/bench-all-gg.txt diff --git a/examples/bench/bench.cpp b/examples/bench/bench.cpp index b77621ac884..cac9385c82f 100644 --- a/examples/bench/bench.cpp +++ b/examples/bench/bench.cpp @@ -12,7 +12,8 @@ struct whisper_params { std::string model = "models/ggml-base.en.bin"; - bool use_gpu = true; + bool use_gpu = true; + bool flash_attn = false; }; void whisper_print_usage(int argc, char ** argv, const whisper_params & params); @@ -25,10 +26,11 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { whisper_print_usage(argc, argv, params); exit(0); } - else if (arg == "-t" || arg == "--threads") { params.n_threads = std::stoi(argv[++i]); } - else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; } - else if (arg == "-w" || arg == "--what") { params.what = atoi(argv[++i]); } - else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-t" || arg == "--threads") { params.n_threads = std::stoi(argv[++i]); } + else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; } + else if (arg == "-w" || arg == "--what") { params.what = atoi(argv[++i]); } + else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else { fprintf(stderr, "error: unknown argument: %s\n", arg.c_str()); whisper_print_usage(argc, argv, params); @@ -49,6 +51,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -m FNAME, --model FNAME [%-7s] model path\n", params.model.c_str()); fprintf(stderr, " -w N, --what N [%-7d] what to benchmark:\n", params.what); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] enable flash attention\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " %-7s 0 - whisper\n", ""); fprintf(stderr, " %-7s 1 - memcpy\n", ""); fprintf(stderr, " %-7s 2 - ggml_mul_mat\n", ""); @@ -59,7 +62,9 @@ int whisper_bench_full(const whisper_params & params) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams); diff --git a/examples/command/command.cpp b/examples/command/command.cpp index ec749d60247..cd6cc023994 100644 --- a/examples/command/command.cpp +++ b/examples/command/command.cpp @@ -44,6 +44,7 @@ struct whisper_params { bool print_energy = false; bool no_timestamps = true; bool use_gpu = true; + bool flash_attn = false; std::string language = "en"; std::string model = "models/ggml-base.en.bin"; @@ -80,6 +81,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-ps" || arg == "--print-special") { params.print_special = true; } else if (arg == "-pe" || arg == "--print-energy") { params.print_energy = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else if (arg == "-l" || arg == "--language") { params.language = argv[++i]; } else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; } else if (arg == "-f" || arg == "--file") { params.fname_out = argv[++i]; } @@ -118,6 +120,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -ps, --print-special [%-7s] print special tokens\n", params.print_special ? "true" : "false"); fprintf(stderr, " -pe, --print-energy [%-7s] print sound energy (for debugging)\n", params.print_energy ? "true" : "false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " -l LANG, --language LANG [%-7s] spoken language\n", params.language.c_str()); fprintf(stderr, " -m FNAME, --model FNAME [%-7s] model path\n", params.model.c_str()); fprintf(stderr, " -f FNAME, --file FNAME [%-7s] text output file name\n", params.fname_out.c_str()); @@ -696,7 +699,9 @@ int main(int argc, char ** argv) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams); diff --git a/examples/lsp/lsp.cpp b/examples/lsp/lsp.cpp index e5f8360f83d..3df54266a25 100644 --- a/examples/lsp/lsp.cpp +++ b/examples/lsp/lsp.cpp @@ -31,6 +31,7 @@ struct whisper_params { bool print_special = false; bool print_energy = false; bool use_gpu = true; + bool flash_attn = false; std::string language = "en"; std::string model = "models/ggml-base.en.bin"; @@ -74,6 +75,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-ps" || arg == "--print-special") { params.print_special = true; } else if (arg == "-pe" || arg == "--print-energy") { params.print_energy = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else if (arg == "-l" || arg == "--language") { params.language = argv[++i]; } else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; } else { @@ -105,6 +107,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -ps, --print-special [%-7s] print special tokens\n", params.print_special ? "true" : "false"); fprintf(stderr, " -pe, --print-energy [%-7s] print sound energy (for debugging)\n", params.print_energy ? "true" : "false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " -l LANG, --language LANG [%-7s] spoken language\n", params.language.c_str()); fprintf(stderr, " -m FNAME, --model FNAME [%-7s] model path\n", params.model.c_str()); fprintf(stderr, "\n"); @@ -436,7 +439,10 @@ int main(int argc, char ** argv) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; + struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams); // init audio diff --git a/examples/main/main.cpp b/examples/main/main.cpp index d11c1c3f81b..45eb17fe7f3 100644 --- a/examples/main/main.cpp +++ b/examples/main/main.cpp @@ -70,6 +70,7 @@ struct whisper_params { bool no_timestamps = false; bool log_score = false; bool use_gpu = true; + bool flash_attn = false; std::string language = "en"; std::string prompt; @@ -168,7 +169,8 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-dtw" || arg == "--dtw") { params.dtw = argv[++i]; } else if (arg == "-ls" || arg == "--log-score") { params.log_score = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } - else if ( arg == "--suppress-regex") { params.suppress_regex = argv[++i]; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } + else if ( arg == "--suppress-regex") { params.suppress_regex = argv[++i]; } else if ( arg == "--grammar") { params.grammar = argv[++i]; } else if ( arg == "--grammar-rule") { params.grammar_rule = argv[++i]; } else if ( arg == "--grammar-penalty") { params.grammar_penalty = std::stof(argv[++i]); } @@ -234,6 +236,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -dtw MODEL --dtw MODEL [%-7s] compute token-level timestamps\n", params.dtw.c_str()); fprintf(stderr, " -ls, --log-score [%-7s] log best decoder scores of tokens\n", params.log_score?"true":"false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " --suppress-regex REGEX [%-7s] regular expression matching tokens to suppress\n", params.suppress_regex.c_str()); fprintf(stderr, " --grammar GRAMMAR [%-7s] GBNF grammar to guide decoding\n", params.grammar.c_str()); fprintf(stderr, " --grammar-rule RULE [%-7s] top-level GBNF grammar rule name\n", params.grammar_rule.c_str()); @@ -977,7 +980,9 @@ int main(int argc, char ** argv) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; if (!params.dtw.empty()) { cparams.dtw_token_timestamps = true; diff --git a/examples/server/server.cpp b/examples/server/server.cpp index e3b96698228..c78b3026e18 100644 --- a/examples/server/server.cpp +++ b/examples/server/server.cpp @@ -75,6 +75,7 @@ struct whisper_params { bool print_progress = false; bool no_timestamps = false; bool use_gpu = true; + bool flash_attn = false; std::string language = "en"; std::string prompt = ""; @@ -178,6 +179,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params, serve else if (arg == "-oved" || arg == "--ov-e-device") { params.openvino_encode_device = argv[++i]; } else if (arg == "-dtw" || arg == "--dtw") { params.dtw = argv[++i]; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } // server params else if ( arg == "--port") { sparams.port = std::stoi(argv[++i]); } else if ( arg == "--host") { sparams.hostname = argv[++i]; } @@ -502,7 +504,10 @@ int main(int argc, char ** argv) { } // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; + if (!params.dtw.empty()) { cparams.dtw_token_timestamps = true; cparams.dtw_aheads_preset = WHISPER_AHEADS_NONE; diff --git a/examples/stream/stream.cpp b/examples/stream/stream.cpp index b82e379dc61..60c1b0894e4 100644 --- a/examples/stream/stream.cpp +++ b/examples/stream/stream.cpp @@ -36,6 +36,7 @@ struct whisper_params { bool tinydiarize = false; bool save_audio = false; // save audio to wav file bool use_gpu = true; + bool flash_attn = false; std::string language = "en"; std::string model = "models/ggml-base.en.bin"; @@ -72,6 +73,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-tdrz" || arg == "--tinydiarize") { params.tinydiarize = true; } else if (arg == "-sa" || arg == "--save-audio") { params.save_audio = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else { fprintf(stderr, "error: unknown argument: %s\n", arg.c_str()); @@ -109,6 +111,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -tdrz, --tinydiarize [%-7s] enable tinydiarize (requires a tdrz model)\n", params.tinydiarize ? "true" : "false"); fprintf(stderr, " -sa, --save-audio [%-7s] save the recorded audio to a file\n", params.save_audio ? "true" : "false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU inference\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention during inference\n", params.flash_attn ? "true" : "false"); fprintf(stderr, "\n"); } @@ -153,7 +156,9 @@ int main(int argc, char ** argv) { } struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams); diff --git a/examples/talk-llama/talk-llama.cpp b/examples/talk-llama/talk-llama.cpp index 838d6f56357..4aab62b9a6f 100644 --- a/examples/talk-llama/talk-llama.cpp +++ b/examples/talk-llama/talk-llama.cpp @@ -66,6 +66,7 @@ struct whisper_params { bool no_timestamps = true; bool verbose_prompt = false; bool use_gpu = true; + bool flash_attn = false; std::string person = "Georgi"; std::string bot_name = "LLaMA"; @@ -105,6 +106,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-pe" || arg == "--print-energy") { params.print_energy = true; } else if (arg == "-vp" || arg == "--verbose-prompt") { params.verbose_prompt = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else if (arg == "-p" || arg == "--person") { params.person = argv[++i]; } else if (arg == "-bn" || arg == "--bot-name") { params.bot_name = argv[++i]; } else if (arg == "--session") { params.path_session = argv[++i]; } @@ -123,7 +125,6 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { } } else if (arg == "-f" || arg == "--file") { params.fname_out = argv[++i]; } - else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } else { fprintf(stderr, "error: unknown argument: %s\n", arg.c_str()); whisper_print_usage(argc, argv, params); @@ -154,6 +155,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -pe, --print-energy [%-7s] print sound energy (for debugging)\n", params.print_energy ? "true" : "false"); fprintf(stderr, " -vp, --verbose-prompt [%-7s] print prompt at start\n", params.verbose_prompt ? "true" : "false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " -p NAME, --person NAME [%-7s] person name (for prompt selection)\n", params.person.c_str()); fprintf(stderr, " -bn NAME, --bot-name NAME [%-7s] bot name (to display)\n", params.bot_name.c_str()); fprintf(stderr, " -w TEXT, --wake-command T [%-7s] wake-up command to listen for\n", params.wake_cmd.c_str()); @@ -285,7 +287,9 @@ int main(int argc, char ** argv) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx_wsp = whisper_init_from_file_with_params(params.model_wsp.c_str(), cparams); if (!ctx_wsp) { @@ -316,6 +320,7 @@ int main(int argc, char ** argv) { lcparams.n_ctx = 2048; lcparams.seed = 1; lcparams.n_threads = params.n_threads; + lcparams.flash_attn = params.flash_attn; struct llama_context * ctx_llama = llama_new_context_with_model(model_llama, lcparams); diff --git a/examples/talk/talk.cpp b/examples/talk/talk.cpp index c1c6f8ba0b2..3e34e5724ff 100644 --- a/examples/talk/talk.cpp +++ b/examples/talk/talk.cpp @@ -32,6 +32,7 @@ struct whisper_params { bool print_energy = false; bool no_timestamps = true; bool use_gpu = true; + bool flash_attn = false; std::string person = "Santa"; std::string language = "en"; @@ -64,6 +65,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-ps" || arg == "--print-special") { params.print_special = true; } else if (arg == "-pe" || arg == "--print-energy") { params.print_energy = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else if (arg == "-p" || arg == "--person") { params.person = argv[++i]; } else if (arg == "-l" || arg == "--language") { params.language = argv[++i]; } else if (arg == "-mw" || arg == "--model-whisper") { params.model_wsp = argv[++i]; } @@ -99,6 +101,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -ps, --print-special [%-7s] print special tokens\n", params.print_special ? "true" : "false"); fprintf(stderr, " -pe, --print-energy [%-7s] print sound energy (for debugging)\n", params.print_energy ? "true" : "false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " -p NAME, --person NAME [%-7s] person name (for prompt selection)\n", params.person.c_str()); fprintf(stderr, " -l LANG, --language LANG [%-7s] spoken language\n", params.language.c_str()); fprintf(stderr, " -mw FILE, --model-whisper [%-7s] whisper model file\n", params.model_wsp.c_str()); @@ -188,7 +191,9 @@ int main(int argc, char ** argv) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx_wsp = whisper_init_from_file_with_params(params.model_wsp.c_str(), cparams); diff --git a/examples/wchess/wchess.cmd/wchess.cmd.cpp b/examples/wchess/wchess.cmd/wchess.cmd.cpp index f66b1765f5b..09e53f13172 100644 --- a/examples/wchess/wchess.cmd/wchess.cmd.cpp +++ b/examples/wchess/wchess.cmd/wchess.cmd.cpp @@ -32,6 +32,7 @@ struct whisper_params { bool print_energy = false; bool no_timestamps = true; bool use_gpu = true; + bool flash_attn = false; std::string language = "en"; std::string model = "models/ggml-base.en.bin"; @@ -61,6 +62,7 @@ void whisper_print_usage(int /*argc*/, char ** argv, const whisper_params & para fprintf(stderr, " -ps, --print-special [%-7s] print special tokens\n", params.print_special ? "true" : "false"); fprintf(stderr, " -pe, --print-energy [%-7s] print sound energy (for debugging)\n", params.print_energy ? "true" : "false"); fprintf(stderr, " -ng, --no-gpu [%-7s] disable GPU\n", params.use_gpu ? "false" : "true"); + fprintf(stderr, " -fa, --flash-attn [%-7s] flash attention during decoding\n", params.flash_attn ? "true" : "false"); fprintf(stderr, " -l LANG, --language LANG [%-7s] spoken language\n", params.language.c_str()); fprintf(stderr, " -m FNAME, --model FNAME [%-7s] model path\n", params.model.c_str()); fprintf(stderr, " -f FNAME, --file FNAME [%-7s] text output file name\n", params.fname_out.c_str()); @@ -92,6 +94,7 @@ bool whisper_params_parse(int argc, char ** argv, whisper_params & params) { else if (arg == "-ps" || arg == "--print-special") { params.print_special = true; } else if (arg == "-pe" || arg == "--print-energy") { params.print_energy = true; } else if (arg == "-ng" || arg == "--no-gpu") { params.use_gpu = false; } + else if (arg == "-fa" || arg == "--flash-attn") { params.flash_attn = true; } else if (arg == "-l" || arg == "--language") { params.language = argv[++i]; } else if (arg == "-m" || arg == "--model") { params.model = argv[++i]; } else if (arg == "-f" || arg == "--file") { params.fname_out = argv[++i]; } @@ -183,7 +186,9 @@ int main(int argc, char ** argv) { // whisper init struct whisper_context_params cparams = whisper_context_default_params(); - cparams.use_gpu = params.use_gpu; + + cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams); if (!ctx) { diff --git a/scripts/bench-all-gg.txt b/scripts/bench-all-gg.txt new file mode 100644 index 00000000000..6fd5605a2bd --- /dev/null +++ b/scripts/bench-all-gg.txt @@ -0,0 +1,298 @@ +## M1 Pro + +make -j && ./scripts/bench-all.sh 8 + +Running memcpy benchmark + +memcpy: 39.10 GB/s (heat-up) +memcpy: 44.75 GB/s ( 1 thread) +memcpy: 44.78 GB/s ( 1 thread) +memcpy: 44.97 GB/s ( 2 thread) +memcpy: 48.04 GB/s ( 3 thread) +memcpy: 50.55 GB/s ( 4 thread) +memcpy: 55.20 GB/s ( 5 thread) +memcpy: 65.60 GB/s ( 6 thread) +memcpy: 70.64 GB/s ( 7 thread) +memcpy: 73.34 GB/s ( 8 thread) +sum: -5120002535.000000 + + +make -j && ./scripts/bench-all.sh 1 0 0 + +Running ggml_mul_mat benchmark with 1 threads + + 64 x 64: Q4_0 237.1 GFLOPS (128 runs) | Q4_1 168.6 GFLOPS (128 runs) + 64 x 64: Q5_0 136.4 GFLOPS (128 runs) | Q5_1 135.6 GFLOPS (128 runs) | Q8_0 243.1 GFLOPS (128 runs) + 64 x 64: F16 140.4 GFLOPS (128 runs) | F32 316.6 GFLOPS (128 runs) + 128 x 128: Q4_0 496.6 GFLOPS (128 runs) | Q4_1 348.6 GFLOPS (128 runs) + 128 x 128: Q5_0 273.2 GFLOPS (128 runs) | Q5_1 274.1 GFLOPS (128 runs) | Q8_0 505.1 GFLOPS (128 runs) + 128 x 128: F16 300.4 GFLOPS (128 runs) | F32 653.9 GFLOPS (128 runs) + 256 x 256: Q4_0 791.7 GFLOPS (128 runs) | Q4_1 615.3 GFLOPS (128 runs) + 256 x 256: Q5_0 651.0 GFLOPS (128 runs) | Q5_1 674.7 GFLOPS (128 runs) | Q8_0 803.1 GFLOPS (128 runs) + 256 x 256: F16 869.6 GFLOPS (128 runs) | F32 957.2 GFLOPS (128 runs) + 512 x 512: Q4_0 973.3 GFLOPS (128 runs) | Q4_1 897.9 GFLOPS (128 runs) + 512 x 512: Q5_0 1078.8 GFLOPS (128 runs) | Q5_1 998.4 GFLOPS (128 runs) | Q8_0 752.4 GFLOPS (128 runs) + 512 x 512: F16 892.5 GFLOPS (128 runs) | F32 1399.6 GFLOPS (128 runs) +1024 x 1024: Q4_0 1402.7 GFLOPS (128 runs) | Q4_1 1218.5 GFLOPS (128 runs) +1024 x 1024: Q5_0 1444.8 GFLOPS (128 runs) | Q5_1 1444.7 GFLOPS (128 runs) | Q8_0 1395.7 GFLOPS (128 runs) +1024 x 1024: F16 1524.1 GFLOPS (128 runs) | F32 1726.6 GFLOPS (128 runs) +2048 x 2048: Q4_0 1479.4 GFLOPS ( 87 runs) | Q4_1 1378.5 GFLOPS ( 81 runs) +2048 x 2048: Q5_0 1454.6 GFLOPS ( 85 runs) | Q5_1 1462.9 GFLOPS ( 86 runs) | Q8_0 1483.2 GFLOPS ( 87 runs) +2048 x 2048: F16 1488.0 GFLOPS ( 87 runs) | F32 1538.2 GFLOPS ( 90 runs) +4096 x 4096: Q4_0 1509.7 GFLOPS ( 11 runs) | Q4_1 1433.0 GFLOPS ( 11 runs) +4096 x 4096: Q5_0 1422.4 GFLOPS ( 11 runs) | Q5_1 1437.0 GFLOPS ( 11 runs) | Q8_0 1523.0 GFLOPS ( 12 runs) +4096 x 4096: F16 1551.3 GFLOPS ( 12 runs) | F32 1451.0 GFLOPS ( 11 runs) + +| CPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| M1 Pro | METAL | tiny | 1 | 0 | 39.21 | 1.74 | 0.61 | 0.04 | 22c96b4 | +| M1 Pro | METAL | base | 1 | 0 | 70.76 | 2.60 | 0.93 | 0.06 | 22c96b4 | +| M1 Pro | METAL | small | 1 | 0 | 217.28 | 6.42 | 2.14 | 0.17 | 22c96b4 | +| M1 Pro | METAL | medium | 1 | 0 | 596.74 | 14.43 | 4.75 | 0.45 | 22c96b4 | + + +make -j && ./scripts/bench-all.sh 1 1 1 + +| CPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| M1 Pro | METAL | tiny | 1 | 1 | 30.77 | 1.59 | 0.54 | 0.03 | 22c96b4 | +| M1 Pro | METAL | base | 1 | 1 | 60.42 | 2.29 | 0.81 | 0.05 | 22c96b4 | +| M1 Pro | METAL | small | 1 | 1 | 183.82 | 5.12 | 1.81 | 0.14 | 22c96b4 | +| M1 Pro | METAL | medium | 1 | 1 | 517.92 | 11.60 | 4.01 | 0.38 | 22c96b4 | + + +## M2 Ultra + +make -j && ./scripts/bench-all.sh 8 + +Running memcpy benchmark + +memcpy: 46.58 GB/s (heat-up) +memcpy: 54.16 GB/s ( 1 thread) +memcpy: 54.23 GB/s ( 1 thread) +memcpy: 99.63 GB/s ( 2 thread) +memcpy: 140.59 GB/s ( 3 thread) +memcpy: 176.52 GB/s ( 4 thread) +memcpy: 158.90 GB/s ( 5 thread) +memcpy: 163.00 GB/s ( 6 thread) +memcpy: 189.69 GB/s ( 7 thread) +memcpy: 197.15 GB/s ( 8 thread) +sum: -5120002007.000000 + + +make -j && ./scripts/bench-all.sh 1 + +Running ggml_mul_mat benchmark with 1 threads + + 64 x 64: Q4_0 245.8 GFLOPS (128 runs) | Q4_1 168.6 GFLOPS (128 runs) + 64 x 64: Q5_0 115.7 GFLOPS (128 runs) | Q5_1 125.9 GFLOPS (128 runs) | Q8_0 215.8 GFLOPS (128 runs) + 64 x 64: F16 139.5 GFLOPS (128 runs) | F32 337.2 GFLOPS (128 runs) + 128 x 128: Q4_0 494.8 GFLOPS (128 runs) | Q4_1 350.4 GFLOPS (128 runs) + 128 x 128: Q5_0 257.1 GFLOPS (128 runs) | Q5_1 261.4 GFLOPS (128 runs) | Q8_0 509.4 GFLOPS (128 runs) + 128 x 128: F16 302.3 GFLOPS (128 runs) | F32 672.8 GFLOPS (128 runs) + 256 x 256: Q4_0 795.7 GFLOPS (128 runs) | Q4_1 663.7 GFLOPS (128 runs) + 256 x 256: Q5_0 737.8 GFLOPS (128 runs) | Q5_1 757.6 GFLOPS (128 runs) | Q8_0 827.7 GFLOPS (128 runs) + 256 x 256: F16 872.6 GFLOPS (128 runs) | F32 956.3 GFLOPS (128 runs) + 512 x 512: Q4_0 1188.0 GFLOPS (128 runs) | Q4_1 1085.0 GFLOPS (128 runs) + 512 x 512: Q5_0 1421.1 GFLOPS (128 runs) | Q5_1 1454.9 GFLOPS (128 runs) | Q8_0 1191.4 GFLOPS (128 runs) + 512 x 512: F16 1577.4 GFLOPS (128 runs) | F32 1982.0 GFLOPS (128 runs) +1024 x 1024: Q4_0 2342.6 GFLOPS (128 runs) | Q4_1 1955.8 GFLOPS (128 runs) +1024 x 1024: Q5_0 2306.7 GFLOPS (128 runs) | Q5_1 2217.0 GFLOPS (128 runs) | Q8_0 2230.7 GFLOPS (128 runs) +1024 x 1024: F16 2593.8 GFLOPS (128 runs) | F32 3269.0 GFLOPS (128 runs) +2048 x 2048: Q4_0 3735.7 GFLOPS (128 runs) | Q4_1 3205.3 GFLOPS (128 runs) +2048 x 2048: Q5_0 3584.5 GFLOPS (128 runs) | Q5_1 3621.7 GFLOPS (128 runs) | Q8_0 3622.3 GFLOPS (128 runs) +2048 x 2048: F16 3763.6 GFLOPS (128 runs) | F32 4153.3 GFLOPS (128 runs) +4096 x 4096: Q4_0 3891.1 GFLOPS ( 29 runs) | Q4_1 3554.0 GFLOPS ( 26 runs) +4096 x 4096: Q5_0 3753.1 GFLOPS ( 28 runs) | Q5_1 3750.1 GFLOPS ( 28 runs) | Q8_0 3768.5 GFLOPS ( 28 runs) +4096 x 4096: F16 3864.2 GFLOPS ( 29 runs) | F32 3970.5 GFLOPS ( 29 runs) + + +make -j && ./scripts/bench-all.sh 1 1 0 + +| CPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| M2 ULTRA | METAL | tiny | 1 | 0 | 12.32 | 1.35 | 0.49 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | tiny-q5_0 | 1 | 0 | 11.65 | 1.30 | 0.51 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | tiny-q5_1 | 1 | 0 | 12.08 | 1.30 | 0.51 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | base | 1 | 0 | 17.58 | 1.90 | 0.76 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | base-q5_0 | 1 | 0 | 18.89 | 1.86 | 0.79 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | base-q5_1 | 1 | 0 | 20.69 | 1.88 | 0.79 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | small | 1 | 0 | 49.32 | 3.85 | 1.71 | 0.05 | 22c96b4 | +| M2 ULTRA | METAL | small-q5_0 | 1 | 0 | 54.91 | 3.81 | 1.82 | 0.06 | 22c96b4 | +| M2 ULTRA | METAL | small-q5_1 | 1 | 0 | 54.92 | 3.81 | 1.79 | 0.06 | 22c96b4 | +| M2 ULTRA | METAL | medium | 1 | 0 | 134.34 | 8.04 | 3.82 | 0.13 | 22c96b4 | +| M2 ULTRA | METAL | medium-q5_0 | 1 | 0 | 151.68 | 7.59 | 4.07 | 0.14 | 22c96b4 | +| M2 ULTRA | METAL | medium-q5_1 | 1 | 0 | 151.58 | 7.67 | 4.07 | 0.14 | 22c96b4 | +| M2 ULTRA | METAL | medium-dis | 1 | 0 | 120.82 | 1.07 | 0.41 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | large-v2 | 1 | 0 | 235.63 | 12.27 | 5.85 | 0.22 | 22c96b4 | +| M2 ULTRA | METAL | large-v2-q5_0 | 1 | 0 | 273.38 | 11.17 | 6.40 | 0.26 | 22c96b4 | +| M2 ULTRA | METAL | large-v2-q5_1 | 1 | 0 | 272.44 | 11.32 | 6.29 | 0.26 | 22c96b4 | +| M2 ULTRA | METAL | large-v2-dis | 1 | 0 | 212.51 | 1.20 | 0.47 | 0.02 | 22c96b4 | + + +make -j && ./scripts/bench-all.sh 1 1 1 + +| CPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| M2 ULTRA | METAL | tiny | 1 | 1 | 9.07 | 1.33 | 0.45 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | tiny-q5_0 | 1 | 1 | 9.74 | 1.33 | 0.47 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | tiny-q5_1 | 1 | 1 | 8.93 | 1.31 | 0.46 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | base | 1 | 1 | 15.75 | 1.87 | 0.71 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | base-q5_0 | 1 | 1 | 17.04 | 1.83 | 0.74 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | base-q5_1 | 1 | 1 | 17.17 | 1.83 | 0.74 | 0.02 | 22c96b4 | +| M2 ULTRA | METAL | small | 1 | 1 | 42.33 | 3.64 | 1.60 | 0.05 | 22c96b4 | +| M2 ULTRA | METAL | small-q5_0 | 1 | 1 | 47.61 | 3.63 | 1.70 | 0.05 | 22c96b4 | +| M2 ULTRA | METAL | small-q5_1 | 1 | 1 | 47.70 | 3.66 | 1.68 | 0.05 | 22c96b4 | +| M2 ULTRA | METAL | medium | 1 | 1 | 114.42 | 7.53 | 3.55 | 0.11 | 22c96b4 | +| M2 ULTRA | METAL | medium-q5_0 | 1 | 1 | 132.63 | 7.02 | 3.77 | 0.13 | 22c96b4 | +| M2 ULTRA | METAL | medium-q5_1 | 1 | 1 | 132.28 | 7.10 | 3.76 | 0.13 | 22c96b4 | +| M2 ULTRA | METAL | medium-dis | 1 | 1 | 102.34 | 1.01 | 0.42 | 0.01 | 22c96b4 | +| M2 ULTRA | METAL | large-v2 | 1 | 1 | 203.01 | 11.03 | 5.45 | 0.20 | 22c96b4 | +| M2 ULTRA | METAL | large-v2-q5_0 | 1 | 1 | 240.05 | 10.18 | 5.98 | 0.23 | 22c96b4 | +| M2 ULTRA | METAL | large-v2-q5_1 | 1 | 1 | 239.22 | 10.23 | 5.87 | 0.23 | 22c96b4 | +| M2 ULTRA | METAL | large-v2-dis | 1 | 1 | 181.14 | 1.14 | 0.48 | 0.02 | 22c96b4 | + + + +## Ryzen 9 5950X + RTX 2060 + +make -j && ./scripts/bench-all.sh 8 0 0 + +Running memcpy benchmark + +memcpy: 12.36 GB/s (heat-up) +memcpy: 12.33 GB/s ( 1 thread) +memcpy: 12.38 GB/s ( 1 thread) +memcpy: 14.48 GB/s ( 2 thread) +memcpy: 15.00 GB/s ( 3 thread) +memcpy: 14.77 GB/s ( 4 thread) +memcpy: 14.60 GB/s ( 5 thread) +memcpy: 14.57 GB/s ( 6 thread) +memcpy: 14.34 GB/s ( 7 thread) +memcpy: 14.40 GB/s ( 8 thread) +sum: -5119998076.000000 + +Running ggml_mul_mat benchmark with 8 threads + + 64 x 64: Q4_0 3.1 GFLOPS (128 runs) | Q4_1 3.1 GFLOPS (128 runs) + 64 x 64: Q5_0 3.0 GFLOPS (128 runs) | Q5_1 2.9 GFLOPS (128 runs) | Q8_0 3.1 GFLOPS (128 runs) + 64 x 64: F16 3.0 GFLOPS (128 runs) | F32 3.0 GFLOPS (128 runs) + 128 x 128: Q4_0 21.1 GFLOPS (128 runs) | Q4_1 20.3 GFLOPS (128 runs) + 128 x 128: Q5_0 20.6 GFLOPS (128 runs) | Q5_1 20.4 GFLOPS (128 runs) | Q8_0 22.1 GFLOPS (128 runs) + 128 x 128: F16 21.7 GFLOPS (128 runs) | F32 21.7 GFLOPS (128 runs) + 256 x 256: Q4_0 105.7 GFLOPS (128 runs) | Q4_1 94.4 GFLOPS (128 runs) + 256 x 256: Q5_0 94.8 GFLOPS (128 runs) | Q5_1 87.5 GFLOPS (128 runs) | Q8_0 107.2 GFLOPS (128 runs) + 256 x 256: F16 95.1 GFLOPS (128 runs) | F32 94.3 GFLOPS (128 runs) + 512 x 512: Q4_0 214.7 GFLOPS (128 runs) | Q4_1 189.8 GFLOPS (128 runs) + 512 x 512: Q5_0 187.7 GFLOPS (128 runs) | Q5_1 176.2 GFLOPS (128 runs) | Q8_0 252.2 GFLOPS (128 runs) + 512 x 512: F16 220.8 GFLOPS (128 runs) | F32 218.3 GFLOPS (128 runs) +1024 x 1024: Q4_0 333.7 GFLOPS (128 runs) | Q4_1 305.8 GFLOPS (128 runs) +1024 x 1024: Q5_0 283.2 GFLOPS (128 runs) | Q5_1 268.2 GFLOPS (125 runs) | Q8_0 394.1 GFLOPS (128 runs) +1024 x 1024: F16 355.0 GFLOPS (128 runs) | F32 313.0 GFLOPS (128 runs) +2048 x 2048: Q4_0 395.0 GFLOPS ( 23 runs) | Q4_1 380.6 GFLOPS ( 23 runs) +2048 x 2048: Q5_0 336.6 GFLOPS ( 20 runs) | Q5_1 318.4 GFLOPS ( 19 runs) | Q8_0 482.6 GFLOPS ( 29 runs) +2048 x 2048: F16 424.5 GFLOPS ( 25 runs) | F32 337.7 GFLOPS ( 20 runs) +4096 x 4096: Q4_0 412.8 GFLOPS ( 4 runs) | Q4_1 405.1 GFLOPS ( 3 runs) +4096 x 4096: Q5_0 346.0 GFLOPS ( 3 runs) | Q5_1 334.6 GFLOPS ( 3 runs) | Q8_0 502.6 GFLOPS ( 4 runs) +4096 x 4096: F16 412.5 GFLOPS ( 4 runs) | F32 274.0 GFLOPS ( 3 runs) + +| CPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| Ryzen 9 5950X | AVX2 | tiny | 8 | 0 | 195.29 | 1.57 | 0.51 | 0.26 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | tiny-q5_0 | 8 | 0 | 213.33 | 1.10 | 0.50 | 0.30 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | tiny-q5_1 | 8 | 0 | 219.38 | 1.18 | 0.53 | 0.32 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | base | 8 | 0 | 424.85 | 3.71 | 1.03 | 0.46 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | base-q5_0 | 8 | 0 | 473.61 | 1.81 | 0.82 | 0.52 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | base-q5_1 | 8 | 0 | 484.14 | 1.92 | 0.85 | 0.56 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | small | 8 | 0 | 1458.32 | 12.66 | 3.09 | 1.26 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | small-q5_0 | 8 | 0 | 1673.22 | 6.42 | 2.18 | 1.45 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | small-q5_1 | 8 | 0 | 1724.78 | 6.72 | 2.32 | 1.52 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | medium | 8 | 0 | 4333.87 | 36.80 | 8.56 | 3.37 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | medium-q5_0 | 8 | 0 | 5194.09 | 19.21 | 5.71 | 3.97 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | medium-q5_1 | 8 | 0 | 5450.39 | 20.01 | 5.99 | 4.17 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | medium-dis | 8 | 0 | 3995.19 | 5.08 | 1.21 | 0.55 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | large-v2 | 8 | 0 | 8056.16 | 69.74 | 16.11 | 6.13 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | large-v2-q5_0 | 8 | 0 | 9799.58 | 35.16 | 10.49 | 7.28 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | large-v2-q5_1 | 8 | 0 | ms | 36.74 | 11.02 | 7.65 | 22c96b4 | +| Ryzen 9 5950X | AVX2 | large-v2-dis | 8 | 0 | 7490.03 | 7.40 | 1.70 | 0.72 | 22c96b4 | + + +WHISPER_CUDA=1 make -j && ./scripts/bench-all.sh 8 1 0 + +| GPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| RTX 2060 | AVX2 CUDA | tiny | 8 | 0 | 12.54 | 0.93 | 0.29 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | tiny-q5_0 | 8 | 0 | 12.73 | 0.98 | 0.24 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | tiny-q5_1 | 8 | 0 | 12.72 | 0.99 | 0.24 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | base | 8 | 0 | 24.14 | 1.28 | 0.41 | 0.03 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | base-q5_0 | 8 | 0 | 24.58 | 1.38 | 0.35 | 0.03 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | base-q5_1 | 8 | 0 | 24.58 | 1.37 | 0.35 | 0.03 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | small | 8 | 0 | 74.70 | 2.91 | 0.84 | 0.07 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | small-q5_0 | 8 | 0 | 76.12 | 2.84 | 0.77 | 0.08 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | small-q5_1 | 8 | 0 | 76.14 | 2.84 | 0.76 | 0.08 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium | 8 | 0 | 200.69 | 6.46 | 1.83 | 0.17 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium-q5_0 | 8 | 0 | 204.80 | 5.90 | 1.65 | 0.19 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium-q5_1 | 8 | 0 | 205.61 | 5.85 | 1.61 | 0.19 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium-dis | 8 | 0 | 186.17 | 0.86 | 0.24 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2 | 8 | 0 | 347.22 | 10.36 | 2.82 | 0.29 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2-q5_0 | 8 | 0 | 357.06 | 8.81 | 2.58 | 0.34 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2-q5_1 | 8 | 0 | 356.97 | 8.62 | 2.49 | 0.33 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2-dis | 8 | 0 | 318.05 | 1.03 | 0.34 | 0.04 | 22c96b4 | + + +WHISPER_CUDA=1 make -j && ./scripts/bench-all.sh 8 1 1 + +| GPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| RTX 2060 | AVX2 CUDA | tiny | 8 | 1 | 7.21 | 0.76 | 0.29 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | tiny-q5_0 | 8 | 1 | 7.42 | 0.82 | 0.18 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | tiny-q5_1 | 8 | 1 | 7.38 | 0.82 | 0.18 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | base | 8 | 1 | 13.49 | 1.04 | 0.36 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | base-q5_0 | 8 | 1 | 13.94 | 1.13 | 0.26 | 0.03 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | base-q5_1 | 8 | 1 | 13.94 | 1.14 | 0.26 | 0.03 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | small | 8 | 1 | 42.81 | 2.33 | 0.69 | 0.05 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | small-q5_0 | 8 | 1 | 44.43 | 2.25 | 0.59 | 0.06 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | small-q5_1 | 8 | 1 | 44.11 | 2.24 | 0.58 | 0.06 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium | 8 | 1 | 115.47 | 5.17 | 1.45 | 0.11 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium-q5_0 | 8 | 1 | 120.37 | 4.63 | 1.25 | 0.13 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium-q5_1 | 8 | 1 | 120.28 | 4.55 | 1.21 | 0.13 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | medium-dis | 8 | 1 | 101.69 | 0.75 | 0.20 | 0.02 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2 | 8 | 1 | 205.67 | 8.49 | 2.19 | 0.18 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2-q5_0 | 8 | 1 | 214.07 | 6.88 | 1.94 | 0.22 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2-q5_1 | 8 | 1 | 213.98 | 6.70 | 1.86 | 0.22 | 22c96b4 | +| RTX 2060 | AVX2 CUDA | large-v2-dis | 8 | 1 | 176.71 | 0.91 | 0.31 | 0.03 | 22c96b4 | + + + + +# V100 + +WHISPER_CUDA=1 make -j && ./scripts/bench-all.sh 8 1 0 + +| GPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| V100 | AVX2 CUDA | tiny | 1 | 0 | 6.21 | 1.11 | 0.30 | 0.02 | 22c96b4 | +| V100 | AVX2 CUDA | tiny-q5_1 | 1 | 0 | 5.97 | 1.10 | 0.26 | 0.02 | 22c96b4 | +| V100 | AVX2 CUDA | base | 1 | 0 | 10.95 | 1.47 | 0.42 | 0.03 | 22c96b4 | +| V100 | AVX2 CUDA | base-q5_1 | 1 | 0 | 11.13 | 1.53 | 0.36 | 0.03 | 22c96b4 | +| V100 | AVX2 CUDA | small | 1 | 0 | 31.57 | 2.96 | 0.84 | 0.05 | 22c96b4 | +| V100 | AVX2 CUDA | small-q5_1 | 1 | 0 | 32.19 | 3.14 | 0.75 | 0.05 | 22c96b4 | +| V100 | AVX2 CUDA | medium | 1 | 0 | 85.88 | 6.49 | 1.80 | 0.10 | 22c96b4 | +| V100 | AVX2 CUDA | medium-q5_0 | 1 | 0 | 87.53 | 5.82 | 1.37 | 0.10 | 22c96b4 | +| V100 | AVX2 CUDA | large-v2 | 1 | 0 | 142.23 | 8.92 | 2.62 | 0.15 | 22c96b4 | + + +WHISPER_CUDA=1 make -j && ./scripts/bench-all.sh 8 1 1 + +| GPU | Config | Model | Th | FA | Enc. | Dec. | Bch5 | PP | Commit | +| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | +| V100 | AVX2 CUDA | tiny | 1 | 1 | 3.96 | 0.82 | 0.24 | 0.02 | 22c96b4 | +| V100 | AVX2 CUDA | tiny-q5_1 | 1 | 1 | 4.05 | 0.85 | 0.18 | 0.02 | 22c96b4 | +| V100 | AVX2 CUDA | base | 1 | 1 | 7.21 | 1.16 | 0.36 | 0.02 | 22c96b4 | +| V100 | AVX2 CUDA | base-q5_1 | 1 | 1 | 7.39 | 1.21 | 0.26 | 0.02 | 22c96b4 | +| V100 | AVX2 CUDA | small | 1 | 1 | 19.81 | 2.41 | 0.71 | 0.04 | 22c96b4 | +| V100 | AVX2 CUDA | small-q5_1 | 1 | 1 | 20.50 | 2.31 | 0.51 | 0.04 | 22c96b4 | +| V100 | AVX2 CUDA | medium | 1 | 1 | 56.02 | 4.89 | 1.44 | 0.07 | 22c96b4 | +| V100 | AVX2 CUDA | medium-q5_0 | 1 | 1 | 57.85 | 4.73 | 1.09 | 0.08 | 22c96b4 | +| V100 | AVX2 CUDA | large-v2 | 1 | 1 | 92.73 | 7.18 | 2.14 | 0.10 | 22c96b4 | + diff --git a/scripts/bench-all.sh b/scripts/bench-all.sh index 6939dafaca0..8a857c67b6c 100755 --- a/scripts/bench-all.sh +++ b/scripts/bench-all.sh @@ -2,7 +2,7 @@ # Helper script to run the bench tool on all models and print the results in share-able format -printf "Usage: ./bench.sh [n_threads] [encoder-only]\n" +printf "Usage: ./bench.sh [n_threads] [encoder-only] [flash-attn]\n" if [ -z "$1" ]; then n_threads=4 @@ -11,12 +11,19 @@ else fi encoder_only=0 -if [ -z "$2" ]; then +if [ -z "$2" ] || [ "$2" -eq 0 ]; then encoder_only=0 else encoder_only=$2 fi +fattn="" +if [ -z "$3" ] || [ "$3" -eq 0 ]; then + fattn="" +else + fattn="-fa" +fi + models=( \ "tiny" "tiny-q4_0" "tiny-q4_1" "tiny-q5_0" "tiny-q5_1" "tiny-q8_0" \ "base" "base-q4_0" "base-q4_1" "base-q5_0" "base-q5_1" "base-q8_0" \ @@ -44,13 +51,19 @@ if [ "$encoder_only" -eq 0 ]; then printf "\n" fi -printf "| %6s | %6s | %16s | %13s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "CPU" "OS" "Config" "Model" "Th" "Enc." "Dec." "Bch5" "PP" "Commit" -printf "| %6s | %6s | %16s | %13s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "---" "---" "---" "---" "---" "---" "---" "---" "---" "---" +if [ "$fattn" == "-fa" ]; then + fattn_i=1 +else + fattn_i=0 +fi + +printf "| %6s | %6s | %16s | %13s | %3s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "CPU" "OS" "Config" "Model" "Th" "FA" "Enc." "Dec." "Bch5" "PP" "Commit" +printf "| %6s | %6s | %16s | %13s | %3s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "---" "---" "---" "---" "---" "---" "---" "---" "---" "---" "---" for model in "${models[@]}"; do # actual run # store stderr output in a variable in order to parse it later - output=$(./bench -m ./models/ggml-$model.bin -t $n_threads 2>&1) + output=$(./bench -m ./models/ggml-$model.bin -t $n_threads $fattn 2>&1) ret=$? # parse the output: @@ -95,6 +108,6 @@ for model in "${models[@]}"; do commit=$(git rev-parse --short HEAD) if [ $ret -eq 0 ]; then - printf "| | | %16s | %13s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "$config" "$model" "$n_threads" "$encode_time" "$decode_time" "$batchd_time" "$prompt_time" "$commit" + printf "| | | %16s | %13s | %3s | %3s | %7s | %7s | %7s | %7s | %7s |\n" "$config" "$model" "$n_threads" "$fattn_i" "$encode_time" "$decode_time" "$batchd_time" "$prompt_time" "$commit" fi done diff --git a/whisper.cpp b/whisper.cpp index ff4223daf42..84aec8238cd 100644 --- a/whisper.cpp +++ b/whisper.cpp @@ -809,14 +809,15 @@ struct whisper_state { // shared between all decoders whisper_kv_cache kv_cross; + // padded buffer for flash-attention + whisper_kv_cache kv_pad; + whisper_mel mel; whisper_batch batch; whisper_decoder decoders[WHISPER_MAX_DECODERS]; - ggml_backend_t backend = nullptr; - // ggml-alloc: // - stores meta info about the intermediate tensors into the `meta` buffers // - stores the actual tensor data into the `data` buffers @@ -902,14 +903,12 @@ static void read_safe(whisper_model_loader * loader, T & dest) { } static bool kv_cache_init( - const struct whisper_hparams & hparams, struct whisper_kv_cache & cache, ggml_backend_t backend, ggml_type wtype, + int64_t n_text_state, + int64_t n_text_layer, int n_ctx) { - const int64_t n_text_state = hparams.n_text_state; - const int64_t n_text_layer = hparams.n_text_layer; - const int64_t n_mem = n_text_layer*n_ctx; const int64_t n_elements = n_text_state*n_mem; @@ -941,6 +940,8 @@ static bool kv_cache_init( return false; } + ggml_backend_buffer_clear(cache.buffer, 0); + return true; } @@ -1068,6 +1069,26 @@ static void whisper_kv_cache_seq_cp( } } +static uint32_t whisper_kv_cache_get_padding(const struct whisper_context & wctx) { + if (!wctx.params.flash_attn) { + return 1u; + } + +#ifdef GGML_USE_METAL + if (ggml_backend_is_metal(wctx.backend)) { + return 32u; + } +#endif + +#ifdef GGML_USE_CUDA + if (ggml_backend_is_cuda(wctx.backend)) { + return 256u; + } +#endif + + return 1u; +} + // [EXPERIMENTAL] Token-level timestamps with DTW static bool aheads_masks_init( const whisper_context_params & cparams, @@ -1872,6 +1893,14 @@ static struct ggml_cgraph * whisper_build_graph_encoder( const int n_head = hparams.n_audio_head; const int n_layer = hparams.n_audio_layer; + const int n_state_head = n_state/n_head; + + auto & kv_pad = wstate.kv_pad; + + WHISPER_ASSERT(!!kv_pad.ctx); + + const int n_ctx_pad = GGML_PAD(n_ctx, 256); + struct ggml_init_params params = { /*.mem_size =*/ wstate.alloc_encode.meta.size(), /*.mem_buffer =*/ wstate.alloc_encode.meta.data(), @@ -1884,7 +1913,7 @@ static struct ggml_cgraph * whisper_build_graph_encoder( struct ggml_tensor * cur = ggml_view_tensor(ctx0, wstate.embd_conv); - const float KQscale = 1.0f/sqrtf(float(n_state)/n_head); + const float KQscale = 1.0f/sqrtf(float(n_state_head)); // =================================================================== // NOTE: experimenting with partial evaluation of the encoder (ignore) @@ -1934,14 +1963,14 @@ static struct ggml_cgraph * whisper_build_graph_encoder( Qcur = ggml_add(ctx0, Qcur, layer.attn_q_b); - //Qcur = ggml_scale(ctx0, Qcur, pow(float(n_state)/n_head, -0.25)); + //Qcur = ggml_scale(ctx0, Qcur, pow(float(n_state_head), -0.25)); // note: no bias for Key struct ggml_tensor * Kcur = ggml_mul_mat(ctx0, layer.attn_k_w, cur); - //Kcur = ggml_scale(ctx0, Kcur, pow(float(n_state)/n_head, -0.25)); + //Kcur = ggml_scale(ctx0, Kcur, pow(float(n_state_head), -0.25)); struct ggml_tensor * Vcur = ggml_mul_mat(ctx0, layer.attn_v_w, @@ -1955,38 +1984,61 @@ static struct ggml_cgraph * whisper_build_graph_encoder( ggml_permute(ctx0, ggml_cpy(ctx0, Qcur, - ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_state/n_head, n_head, n_ctx)), + ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_state_head, n_head, n_ctx)), 0, 2, 1, 3); - struct ggml_tensor * K = - ggml_permute(ctx0, - ggml_cpy(ctx0, - Kcur, - ggml_new_tensor_3d(ctx0, wctx.itype, n_state/n_head, n_head, n_ctx)), - 0, 2, 1, 3); - - // K * Q - struct ggml_tensor * KQ = ggml_mul_mat(ctx0, K, Q); + if (wctx.params.flash_attn) { + ggml_build_forward_expand(gf, ggml_cpy(ctx0, Kcur, ggml_view_1d(ctx0, kv_pad.k, n_ctx*n_state, 0))); + ggml_build_forward_expand(gf, ggml_cpy(ctx0, Vcur, ggml_view_1d(ctx0, kv_pad.v, n_ctx*n_state, 0))); - struct ggml_tensor * KQ_soft_max = ggml_soft_max_ext(ctx0, KQ, nullptr, KQscale, 0.0f); + struct ggml_tensor * K = + ggml_view_3d(ctx0, kv_pad.k, + n_state_head, n_ctx_pad, n_head, + ggml_element_size(kv_pad.k)*n_state, + ggml_element_size(kv_pad.k)*n_state_head, + 0); - struct ggml_tensor * V = - ggml_cpy(ctx0, - ggml_permute(ctx0, - ggml_reshape_3d(ctx0, - Vcur, - n_state/n_head, n_head, n_ctx), - 1, 2, 0, 3), - ggml_new_tensor_3d(ctx0, wctx.itype, n_ctx, n_state/n_head, n_head) - ); + struct ggml_tensor * V = + ggml_view_3d(ctx0, kv_pad.v, + n_state_head, n_ctx_pad, n_head, + ggml_element_size(kv_pad.v)*n_state, + ggml_element_size(kv_pad.v)*n_state_head, + 0); - struct ggml_tensor * KQV = ggml_mul_mat(ctx0, V, KQ_soft_max); + cur = ggml_flash_attn_ext(ctx0, Q, K, V, nullptr, KQscale, 0.0f); - struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); - - cur = ggml_cpy(ctx0, - KQV_merged, - ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_state, n_ctx)); + cur = ggml_reshape_2d(ctx0, cur, n_state, n_ctx); + } else { + struct ggml_tensor * K = + ggml_permute(ctx0, + ggml_cpy(ctx0, + Kcur, + ggml_new_tensor_3d(ctx0, wctx.itype, n_state_head, n_head, n_ctx)), + 0, 2, 1, 3); + + // K * Q + struct ggml_tensor * KQ = ggml_mul_mat(ctx0, K, Q); + + struct ggml_tensor * KQ_soft_max = ggml_soft_max_ext(ctx0, KQ, nullptr, KQscale, 0.0f); + + struct ggml_tensor * V = + ggml_cpy(ctx0, + ggml_permute(ctx0, + ggml_reshape_3d(ctx0, + Vcur, + n_state_head, n_head, n_ctx), + 1, 2, 0, 3), + ggml_new_tensor_3d(ctx0, wctx.itype, n_ctx, n_state_head, n_head) + ); + + struct ggml_tensor * KQV = ggml_mul_mat(ctx0, V, KQ_soft_max); + + struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); + + cur = ggml_cpy(ctx0, + KQV_merged, + ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_state, n_ctx)); + } } // projection @@ -2085,6 +2137,10 @@ static struct ggml_cgraph * whisper_build_graph_cross( const int n_state = hparams.n_audio_state; const int n_head = hparams.n_audio_head; + const int n_state_head = n_state/n_head; + + const int n_ctx_pad = GGML_PAD(n_ctx, 256); + struct ggml_init_params params = { /*.mem_size =*/ wstate.alloc_cross.meta.size(), /*.mem_buffer =*/ wstate.alloc_cross.meta.data(), @@ -2097,18 +2153,18 @@ static struct ggml_cgraph * whisper_build_graph_cross( struct ggml_tensor * cur = ggml_view_tensor(ctx0, wstate.embd_enc); - const float Kscale = pow(float(n_state) / n_head, -0.25); + const float Kscale = pow(float(n_state_head), -0.25); for (int il = 0; il < model.hparams.n_text_layer; ++il) { auto & layer = model.layers_decoder[il]; - struct ggml_tensor* Kcross = ggml_mul_mat(ctx0, + struct ggml_tensor * Kcross = ggml_mul_mat(ctx0, layer.cross_attn_k_w, cur); Kcross = ggml_scale(ctx0, Kcross, Kscale); - struct ggml_tensor* Vcross = ggml_mul_mat(ctx0, + struct ggml_tensor * Vcross = ggml_mul_mat(ctx0, layer.cross_attn_v_w, cur); @@ -2116,15 +2172,25 @@ static struct ggml_cgraph * whisper_build_graph_cross( Vcross, layer.cross_attn_v_b); - Vcross = ggml_transpose(ctx0, ggml_reshape_2d(ctx0, Vcross, n_state, n_ctx)); + struct ggml_tensor * k; + struct ggml_tensor * v; - struct ggml_tensor * k = ggml_view_1d(ctx0, wstate.kv_cross.k, - n_state*n_ctx, - (ggml_element_size(wstate.kv_cross.k)*n_state)*(il*n_ctx)); + if (wctx.params.flash_attn) { + k = ggml_view_1d(ctx0, wstate.kv_cross.k, n_state*n_ctx, + (ggml_element_size(wstate.kv_cross.k)*n_state)*(il*n_ctx_pad)); - struct ggml_tensor * v = ggml_view_2d(ctx0, wstate.kv_cross.v, n_ctx, n_state, - ( n_ctx)*ggml_element_size(wstate.kv_cross.v), - (il*n_ctx)*ggml_element_size(wstate.kv_cross.v)*n_state); + v = ggml_view_1d(ctx0, wstate.kv_cross.v, n_state*n_ctx, + (ggml_element_size(wstate.kv_cross.v)*n_state)*(il*n_ctx_pad)); + } else { + Vcross = ggml_transpose(ctx0, ggml_reshape_2d(ctx0, Vcross, n_state, n_ctx)); + + k = ggml_view_1d(ctx0, wstate.kv_cross.k, n_state*n_ctx, + (ggml_element_size(wstate.kv_cross.k)*n_state)*(il*n_ctx)); + + v = ggml_view_2d(ctx0, wstate.kv_cross.v, n_ctx, n_state, + ( n_ctx)*ggml_element_size(wstate.kv_cross.v), + (il*n_ctx)*ggml_element_size(wstate.kv_cross.v)*n_state); + } ggml_build_forward_expand(gf, ggml_cpy(ctx0, Kcross, k)); ggml_build_forward_expand(gf, ggml_cpy(ctx0, Vcross, v)); @@ -2195,7 +2261,7 @@ static bool whisper_encode_internal( } if (!whisper_encode_external(wstate)) { - if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { return false; } } else { @@ -2218,7 +2284,7 @@ static bool whisper_encode_internal( return false; } - if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { return false; } } @@ -2234,7 +2300,7 @@ static bool whisper_encode_internal( return false; } - if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { return false; } } @@ -2263,11 +2329,15 @@ static struct ggml_cgraph * whisper_build_graph_decoder( const int n_head = hparams.n_text_head; const int n_layer = hparams.n_text_layer; + const int n_state_head = n_state/n_head; + const int n_tokens = batch.n_tokens; const int n_audio_ctx = wstate.exp_n_audio_ctx > 0 ? wstate.exp_n_audio_ctx : hparams.n_audio_ctx; - const int32_t n_kv = worst_case ? n_ctx : kv_self.n; - const int32_t kv_head = worst_case ? n_ctx - n_tokens : kv_self.head; + const int n_audio_ctx_pad = GGML_PAD(n_audio_ctx, 256); + + const int32_t n_kv = worst_case ? n_ctx : kv_self.n; + const int32_t kv_head = worst_case ? n_ctx - n_tokens : kv_self.head; //WHISPER_LOG_DEBUG("%s: n_past = %d, n_tokens = %d, n_audio_ctx = %d, n_ctx = %d\n", __func__, n_past, n_tokens, n_audio_ctx, n_ctx); @@ -2289,12 +2359,14 @@ static struct ggml_cgraph * whisper_build_graph_decoder( ggml_set_name(position, "position"); ggml_set_input(position); - const float KQscale = pow(float(n_state)/n_head, -0.25); + const float KQscale = pow(float(n_state_head), -0.25); - struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, n_tokens, 1); + struct ggml_tensor * KQ_mask = ggml_new_tensor_3d(ctx0, GGML_TYPE_F32, n_kv, GGML_PAD(n_tokens, GGML_KQ_MASK_PAD), 1); ggml_set_name(KQ_mask, "KQ_mask"); ggml_set_input(KQ_mask); + struct ggml_tensor * KQ_mask_f16 = ggml_cast(ctx0, KQ_mask, GGML_TYPE_F16); + // token encoding + position encoding struct ggml_tensor * cur = ggml_add(ctx0, @@ -2350,12 +2422,25 @@ static struct ggml_cgraph * whisper_build_graph_decoder( Vcur, layer.attn_v_b); - Vcur = ggml_transpose(ctx0, ggml_reshape_2d(ctx0, Vcur, n_state, n_tokens)); + struct ggml_tensor * k; + struct ggml_tensor * v; + + if (wctx.params.flash_attn) { + k = ggml_view_1d(ctx0, kv_self.k, n_tokens*n_state, + (ggml_element_size(kv_self.k)*n_state)*(il*n_ctx + kv_head)); + + v = ggml_view_1d(ctx0, kv_self.v, n_tokens*n_state, + (ggml_element_size(kv_self.v)*n_state)*(il*n_ctx + kv_head)); + } else { + Vcur = ggml_transpose(ctx0, ggml_reshape_2d(ctx0, Vcur, n_state, n_tokens)); - struct ggml_tensor * k = ggml_view_1d(ctx0, kv_self.k, n_tokens*n_state, (ggml_element_size(kv_self.k)*n_state)*(il*n_ctx + kv_head)); - struct ggml_tensor * v = ggml_view_2d(ctx0, kv_self.v, n_tokens, n_state, - ( n_ctx)*ggml_element_size(kv_self.v), - (il*n_ctx)*ggml_element_size(kv_self.v)*n_state + kv_head*ggml_element_size(kv_self.v)); + k = ggml_view_1d(ctx0, kv_self.k, n_tokens*n_state, + (ggml_element_size(kv_self.k)*n_state)*(il*n_ctx + kv_head)); + + v = ggml_view_2d(ctx0, kv_self.v, n_tokens, n_state, + ( n_ctx)*ggml_element_size(kv_self.v), + (il*n_ctx)*ggml_element_size(kv_self.v)*n_state + kv_head*ggml_element_size(kv_self.v)); + } ggml_build_forward_expand(gf, ggml_cpy(ctx0, Kcur, k)); ggml_build_forward_expand(gf, ggml_cpy(ctx0, Vcur, v)); @@ -2365,35 +2450,48 @@ static struct ggml_cgraph * whisper_build_graph_decoder( struct ggml_tensor * Q = ggml_permute(ctx0, - ggml_reshape_3d(ctx0, Qcur, n_state/n_head, n_head, n_tokens), + ggml_reshape_3d(ctx0, Qcur, n_state_head, n_head, n_tokens), 0, 2, 1, 3); struct ggml_tensor * K = ggml_view_3d(ctx0, kv_self.k, - n_state/n_head, n_kv, n_head, + n_state_head, n_kv, n_head, ggml_element_size(kv_self.k)*n_state, - ggml_element_size(kv_self.k)*n_state/n_head, + ggml_element_size(kv_self.k)*n_state_head, ggml_element_size(kv_self.k)*n_state*n_ctx*il); - // K * Q - struct ggml_tensor * KQ = ggml_mul_mat(ctx0, K, Q); + if (wctx.params.flash_attn) { + struct ggml_tensor * V = + ggml_view_3d(ctx0, kv_self.v, + n_state_head, n_kv, n_head, + ggml_element_size(kv_self.v)*n_state, + ggml_element_size(kv_self.v)*n_state_head, + ggml_element_size(kv_self.v)*n_state*n_ctx*il); + + cur = ggml_flash_attn_ext(ctx0, Q, K, V, KQ_mask_f16, 1.0f, 0.0f); + + cur = ggml_reshape_2d(ctx0, cur, n_state, n_tokens); + } else { + // K * Q + struct ggml_tensor * KQ = ggml_mul_mat(ctx0, K, Q); - struct ggml_tensor * KQ_soft_max = ggml_soft_max_ext(ctx0, KQ, KQ_mask, 1.0f, 0.0f); + struct ggml_tensor * KQ_soft_max = ggml_soft_max_ext(ctx0, KQ, KQ_mask, 1.0f, 0.0f); - struct ggml_tensor * V = - ggml_view_3d(ctx0, kv_self.v, - n_kv, n_state/n_head, n_head, - n_ctx*ggml_element_size(kv_self.v), - n_ctx*ggml_element_size(kv_self.v)*n_state/n_head, - n_ctx*ggml_element_size(kv_self.v)*n_state*il); + struct ggml_tensor * V = + ggml_view_3d(ctx0, kv_self.v, + n_kv, n_state_head, n_head, + n_ctx*ggml_element_size(kv_self.v), + n_ctx*ggml_element_size(kv_self.v)*n_state_head, + n_ctx*ggml_element_size(kv_self.v)*n_state*il); - struct ggml_tensor * KQV = ggml_mul_mat(ctx0, V, KQ_soft_max); + struct ggml_tensor * KQV = ggml_mul_mat(ctx0, V, KQ_soft_max); - struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); + struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); - cur = ggml_cpy(ctx0, - KQV_merged, - ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_state, n_tokens)); + cur = ggml_cpy(ctx0, + KQV_merged, + ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_state, n_tokens)); + } } // projection @@ -2432,80 +2530,77 @@ static struct ggml_cgraph * whisper_build_graph_decoder( Qcur, layer.cross_attn_q_b); - Qcur = ggml_scale(ctx0, Qcur, KQscale); - - // Kcross is already scaled - struct ggml_tensor * Kcross = - ggml_view_3d(ctx0, wstate.kv_cross.k, - n_state/n_head, n_audio_ctx, n_head, - ggml_element_size(wstate.kv_cross.k)*n_state, - ggml_element_size(wstate.kv_cross.k)*n_state/n_head, - ggml_element_size(wstate.kv_cross.k)*n_state*n_audio_ctx*il); - - //struct ggml_tensor * Vcross = - // ggml_reshape_3d(ctx0, - // ggml_view_1d(ctx0, wstate.kv_cross.v, n_audio_ctx*n_state, il*n_audio_ctx*ggml_element_size(wstate.kv_cross.v)*n_state), - // n_state/n_head, n_head, n_audio_ctx); - - //struct ggml_tensor * V_trans = - // ggml_cpy(ctx0, - // ggml_permute(ctx0, Vcross, 1, 2, 0, 3), - // ggml_new_tensor_3d(ctx0, Vcross->type, n_audio_ctx, n_state/n_head, n_head)); - - struct ggml_tensor * V = - ggml_view_3d(ctx0, wstate.kv_cross.v, - n_audio_ctx, n_state/n_head, n_head, - n_audio_ctx*ggml_element_size(wstate.kv_cross.v), - n_audio_ctx*ggml_element_size(wstate.kv_cross.v)*n_state/n_head, - n_audio_ctx*ggml_element_size(wstate.kv_cross.v)*n_state*il); - - // ------ - struct ggml_tensor * Q = ggml_permute(ctx0, - ggml_reshape_3d(ctx0, Qcur, n_state/n_head, n_head, n_tokens), + ggml_reshape_3d(ctx0, Qcur, n_state_head, n_head, n_tokens), 0, 2, 1, 3); - // K * Q - struct ggml_tensor * KQ = ggml_mul_mat(ctx0, Kcross, Q); - - //struct ggml_tensor * KQ_scaled = - // ggml_scale(ctx0, - // KQ, - // ggml_new_f32(ctx0, 1.0f/sqrt(float(n_state)/n_head)) - // ); + if (wctx.params.flash_attn) { + struct ggml_tensor * Kcross = + ggml_view_3d(ctx0, wstate.kv_cross.k, + n_state_head, n_audio_ctx_pad, n_head, + ggml_element_size(wstate.kv_cross.k)*n_state, + ggml_element_size(wstate.kv_cross.k)*n_state_head, + ggml_element_size(wstate.kv_cross.k)*n_state*n_audio_ctx_pad*il); - // no masking for cross-attention - //struct ggml_tensor * KQ_masked = ggml_diag_mask_inf(ctx0, KQ_scaled, n_past); + struct ggml_tensor * Vcross = + ggml_view_3d(ctx0, wstate.kv_cross.v, + n_state_head, n_audio_ctx_pad, n_head, + ggml_element_size(wstate.kv_cross.v)*n_state, + ggml_element_size(wstate.kv_cross.v)*n_state_head, + ggml_element_size(wstate.kv_cross.v)*n_state*n_audio_ctx_pad*il); - struct ggml_tensor * KQ_soft_max = ggml_soft_max(ctx0, KQ); + cur = ggml_flash_attn_ext(ctx0, Q, Kcross, Vcross, nullptr, KQscale, 0.0f); - // [EXPERIMENTAL] Token-level timestamps with DTW - if (wctx.params.dtw_token_timestamps) { - if (wstate.aheads_masks.m[il] != nullptr) { - struct ggml_tensor * aheads_KQs = ggml_reshape_2d(ctx0, KQ_soft_max, KQ_soft_max->ne[0] * KQ_soft_max->ne[1], KQ_soft_max->ne[2]); - aheads_KQs = ggml_transpose(ctx0, aheads_KQs); - aheads_KQs = ggml_cont(ctx0, aheads_KQs); - aheads_KQs = ggml_mul_mat(ctx0, wstate.aheads_masks.m[il], aheads_KQs); - aheads_KQs = ggml_transpose(ctx0, aheads_KQs); - aheads_KQs = ggml_cont(ctx0, aheads_KQs); - aheads_KQs = ggml_reshape_3d(ctx0, aheads_KQs, KQ_soft_max->ne[0], KQ_soft_max->ne[1], wstate.aheads_masks.m[il]->ne[1]); - if (aheads_cross_QKs == NULL) { - aheads_cross_QKs = aheads_KQs; - } else { - aheads_cross_QKs = ggml_concat(ctx0, aheads_cross_QKs, aheads_KQs); + cur = ggml_reshape_2d(ctx0, cur, n_state, n_tokens); + } else { + struct ggml_tensor * Kcross = + ggml_view_3d(ctx0, wstate.kv_cross.k, + n_state_head, n_audio_ctx, n_head, + ggml_element_size(wstate.kv_cross.k)*n_state, + ggml_element_size(wstate.kv_cross.k)*n_state_head, + ggml_element_size(wstate.kv_cross.k)*n_state*n_audio_ctx*il); + + struct ggml_tensor * Vcross = + ggml_view_3d(ctx0, wstate.kv_cross.v, + n_audio_ctx, n_state_head, n_head, + n_audio_ctx*ggml_element_size(wstate.kv_cross.v), + n_audio_ctx*ggml_element_size(wstate.kv_cross.v)*n_state_head, + n_audio_ctx*ggml_element_size(wstate.kv_cross.v)*n_state*il); + + // ------ + + // K * Q + struct ggml_tensor * KQ = ggml_mul_mat(ctx0, Kcross, Q); + + struct ggml_tensor * KQ_soft_max = ggml_soft_max_ext(ctx0, KQ, nullptr, KQscale, 0.0f); + + // [EXPERIMENTAL] Token-level timestamps with DTW + if (wctx.params.dtw_token_timestamps) { + if (wstate.aheads_masks.m[il] != nullptr) { + struct ggml_tensor * aheads_KQs = ggml_reshape_2d(ctx0, KQ_soft_max, KQ_soft_max->ne[0] * KQ_soft_max->ne[1], KQ_soft_max->ne[2]); + aheads_KQs = ggml_transpose(ctx0, aheads_KQs); + aheads_KQs = ggml_cont(ctx0, aheads_KQs); + aheads_KQs = ggml_mul_mat(ctx0, wstate.aheads_masks.m[il], aheads_KQs); + aheads_KQs = ggml_transpose(ctx0, aheads_KQs); + aheads_KQs = ggml_cont(ctx0, aheads_KQs); + aheads_KQs = ggml_reshape_3d(ctx0, aheads_KQs, KQ_soft_max->ne[0], KQ_soft_max->ne[1], wstate.aheads_masks.m[il]->ne[1]); + if (aheads_cross_QKs == NULL) { + aheads_cross_QKs = aheads_KQs; + } else { + aheads_cross_QKs = ggml_concat(ctx0, aheads_cross_QKs, aheads_KQs); + } } } - } - struct ggml_tensor * KQV = ggml_mul_mat(ctx0, V, KQ_soft_max); + struct ggml_tensor * KQV = ggml_mul_mat(ctx0, Vcross, KQ_soft_max); - struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); + struct ggml_tensor * KQV_merged = ggml_permute(ctx0, KQV, 0, 2, 1, 3); - // cur = KQV_merged.contiguous().view(n_state, n_tokens) - cur = ggml_cpy(ctx0, - KQV_merged, - ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_state, n_tokens)); + cur = ggml_cpy(ctx0, + KQV_merged, + ggml_new_tensor_2d(ctx0, GGML_TYPE_F32, n_state, n_tokens)); + } } // projection @@ -2638,7 +2733,9 @@ static bool whisper_decode_internal( return false; } - kv_self.n = whisper_kv_cache_cell_max(kv_self); + const uint32_t pad = whisper_kv_cache_get_padding(wctx); + kv_self.n = std::min(kv_self.size, std::max(pad, GGML_PAD(whisper_kv_cache_cell_max(kv_self), pad))); + //kv_self.n = std::min((int32_t) hparams.n_text_ctx, std::max(32, whisper_kv_cache_cell_max(kv_self))); //printf("n_tokens = %5d, kv_self.head = %5d, kv_self.n = %5d, seq_id = %5d\n", batch.n_tokens, kv_self.head, kv_self.n, batch.seq_id[0][0]); } @@ -2672,9 +2769,10 @@ static bool whisper_decode_internal( struct ggml_tensor * KQ_mask = ggml_graph_get_tensor(gf, "KQ_mask"); auto & kv_self = wstate.kv_self; - const int32_t n_kv = kv_self.n; - wstate.inp_mask.resize(n_kv*n_tokens); + const int32_t n_kv = kv_self.n; + + wstate.inp_mask.resize(ggml_nelements(KQ_mask)); float * data = wstate.inp_mask.data(); memset(data, 0, ggml_nbytes(KQ_mask)); @@ -2690,6 +2788,12 @@ static bool whisper_decode_internal( } } } + + for (int i = n_tokens; i < GGML_PAD(n_tokens, GGML_KQ_MASK_PAD); ++i) { + for (int j = 0; j < n_kv; ++j) { + data[h*(n_kv*n_tokens) + i*n_kv + j] = -INFINITY; + } + } } ggml_backend_tensor_set(KQ_mask, wstate.inp_mask.data(), 0, ggml_nelements(KQ_mask)*sizeof(float)); @@ -2697,7 +2801,7 @@ static bool whisper_decode_internal( logits = gf->nodes[gf->n_nodes - 1]; - if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { return false; } } @@ -3144,18 +3248,14 @@ struct whisper_state * whisper_init_state(whisper_context * ctx) { whisper_state * state = new whisper_state; - state->backend = whisper_backend_init(ctx->params); - if (!state->backend) { - WHISPER_LOG_ERROR("%s: whisper_backend_init() failed\n", __func__); - whisper_free_state(state); - return nullptr; - } - // at this point, we don't know yet how many decoders will be used, so we overallocate 3x ctx // in theory, there can be a case where this is not enough, but in practice it should always be enough const int factor = 3; - if (!kv_cache_init(ctx->model.hparams, state->kv_self, ctx->backend, ctx->itype, factor*ctx->model.hparams.n_text_ctx)) { + if (!kv_cache_init(state->kv_self, ctx->backend, ctx->itype, + ctx->model.hparams.n_text_state, + ctx->model.hparams.n_text_layer, + GGML_PAD(ctx->model.hparams.n_text_ctx, 256)*factor)) { WHISPER_LOG_ERROR("%s: kv_cache_init() failed for self-attention cache\n", __func__); whisper_free_state(state); return nullptr; @@ -3166,7 +3266,10 @@ struct whisper_state * whisper_init_state(whisper_context * ctx) { WHISPER_LOG_INFO("%s: kv self size = %7.2f MB\n", __func__, memory_size / 1e6); } - if (!kv_cache_init(ctx->model.hparams, state->kv_cross, ctx->backend, ctx->itype, ctx->model.hparams.n_audio_ctx)) { + if (!kv_cache_init(state->kv_cross, ctx->backend, ctx->itype, + ctx->model.hparams.n_text_state, + ctx->model.hparams.n_text_layer, + GGML_PAD(ctx->model.hparams.n_audio_ctx, 256))) { WHISPER_LOG_ERROR("%s: kv_cache_init() failed for cross-attention cache\n", __func__); whisper_free_state(state); return nullptr; @@ -3177,6 +3280,20 @@ struct whisper_state * whisper_init_state(whisper_context * ctx) { WHISPER_LOG_INFO("%s: kv cross size = %7.2f MB\n", __func__, memory_size / 1e6); } + if (!kv_cache_init(state->kv_pad, ctx->backend, ctx->itype, + ctx->model.hparams.n_audio_state, + 1, + GGML_PAD(ctx->model.hparams.n_audio_ctx, 256))) { + WHISPER_LOG_ERROR("%s: kv_cache_init() failed for self-attention cache\n", __func__); + whisper_free_state(state); + return nullptr; + } + + { + const size_t memory_size = ggml_nbytes(state->kv_pad.k) + ggml_nbytes(state->kv_pad.v); + WHISPER_LOG_INFO("%s: kv pad size = %7.2f MB\n", __func__, memory_size / 1e6); + } + // [EXPERIMENTAL] Token-level timestamps with DTW if (ctx->params.dtw_token_timestamps) { if (!aheads_masks_init(ctx->params, ctx->model.hparams, state->aheads_masks, ctx->backend)) { @@ -3347,6 +3464,7 @@ int whisper_ctx_init_openvino_encoder( struct whisper_context_params whisper_context_default_params() { struct whisper_context_params result = { /*.use_gpu =*/ true, + /*.flash_attn =*/ false, /*.gpu_device =*/ 0, /*.dtw_token_timestamps =*/ false, @@ -3445,6 +3563,16 @@ struct whisper_context * whisper_init_from_buffer_with_params_no_state(void * bu struct whisper_context * whisper_init_with_params_no_state(struct whisper_model_loader * loader, struct whisper_context_params params) { ggml_time_init(); + if (params.flash_attn && params.dtw_token_timestamps) { + WHISPER_LOG_WARN("%s: dtw_token_timestamps is not supported with flash_attn - disabling\n", __func__); + params.dtw_token_timestamps = false; + } + + WHISPER_LOG_INFO("%s: use gpu = %d\n", __func__, params.use_gpu); + WHISPER_LOG_INFO("%s: flash attn = %d\n", __func__, params.flash_attn); + WHISPER_LOG_INFO("%s: gpu_device = %d\n", __func__, params.gpu_device); + WHISPER_LOG_INFO("%s: dtw = %d\n", __func__, params.dtw_token_timestamps); + whisper_context * ctx = new whisper_context; ctx->params = params; @@ -3533,6 +3661,7 @@ void whisper_free_state(struct whisper_state * state) { if (state) { kv_cache_free(state->kv_self); kv_cache_free(state->kv_cross); + kv_cache_free(state->kv_pad); #ifdef WHISPER_USE_COREML if (state->ctx_coreml != nullptr) { @@ -3555,8 +3684,6 @@ void whisper_free_state(struct whisper_state * state) { ggml_gallocr_free(state->alloc_cross.alloc); ggml_gallocr_free(state->alloc_decode.alloc); - ggml_backend_free(state->backend); - // [EXPERIMENTAL] Token-level timestamps with DTW aheads_masks_free(state->aheads_masks); diff --git a/whisper.h b/whisper.h index 6a875d3bbb9..9c7c58d874b 100644 --- a/whisper.h +++ b/whisper.h @@ -113,6 +113,7 @@ extern "C" { struct whisper_context_params { bool use_gpu; + bool flash_attn; int gpu_device; // CUDA device // [EXPERIMENTAL] Token-level timestamps with DTW From 08981d1bacbe494ff1c943af6c577c669a2d9f4d Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Wed, 15 May 2024 09:59:48 +0300 Subject: [PATCH 09/17] release : v1.6.0 --- CMakeLists.txt | 2 +- README.md | 2 +- bindings/ios | 2 +- bindings/javascript/package.json | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index cdffbcaa1c0..588aa61cd11 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -3,7 +3,7 @@ cmake_minimum_required (VERSION 3.5) # Allow for the creation of solution folders. set_property(GLOBAL PROPERTY USE_FOLDERS ON) -project(whisper.cpp VERSION 1.5.5) +project(whisper.cpp VERSION 1.6.0) set(SOVERSION 1) # Add path to modules diff --git a/README.md b/README.md index 33570ef02bc..0c34e8dffce 100644 --- a/README.md +++ b/README.md @@ -6,7 +6,7 @@ [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT) [![npm](https://img.shields.io/npm/v/whisper.cpp.svg)](https://www.npmjs.com/package/whisper.cpp/) -Stable: [v1.5.5](https://github.com/ggerganov/whisper.cpp/releases/tag/v1.5.5) / [Roadmap | F.A.Q.](https://github.com/ggerganov/whisper.cpp/discussions/126) +Stable: [v1.6.0](https://github.com/ggerganov/whisper.cpp/releases/tag/v1.6.0) / [Roadmap | F.A.Q.](https://github.com/ggerganov/whisper.cpp/discussions/126) High-performance inference of [OpenAI's Whisper](https://github.com/openai/whisper) automatic speech recognition (ASR) model: diff --git a/bindings/ios b/bindings/ios index 0c6cfa58a2c..5cfcfb0801b 160000 --- a/bindings/ios +++ b/bindings/ios @@ -1 +1 @@ -Subproject commit 0c6cfa58a2c7384f567a5680459a0deb79224881 +Subproject commit 5cfcfb0801be756d8347822b472e4b5e343f403f diff --git a/bindings/javascript/package.json b/bindings/javascript/package.json index f64d975663e..354d0ce903c 100644 --- a/bindings/javascript/package.json +++ b/bindings/javascript/package.json @@ -1,6 +1,6 @@ { "name": "whisper.cpp", - "version": "1.5.5", + "version": "1.6.0", "description": "Whisper speech recognition", "main": "whisper.js", "scripts": { From 4798be1f9a8e9bb4aaf05884e852902274235fdc Mon Sep 17 00:00:00 2001 From: Tamotsu Takahashi Date: Sun, 19 May 2024 17:49:26 +0900 Subject: [PATCH 10/17] ci: Update build.yml to suppress warnings about node.js versions (#2166) * Update actions to suppress warnings about old node.js https://github.blog/changelog/2023-09-22-github-actions-transitioning-from-node-16-to-node-20/ * Update actions/upload-artifact, specify android cmdline-tools-version * Use java 20 gradle 8.1 complains against 21 https://docs.gradle.org/current/userguide/compatibility.html --- .github/workflows/build.yml | 92 ++++++++++++++++++------------------- 1 file changed, 46 insertions(+), 46 deletions(-) diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml index 2d75fc31466..e9bf9c28292 100644 --- a/.github/workflows/build.yml +++ b/.github/workflows/build.yml @@ -15,10 +15,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Set up QEMU - uses: docker/setup-qemu-action@v2 + uses: docker/setup-qemu-action@v3 - name: Build ${{ matrix.arch }} run: | @@ -36,7 +36,7 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Dependencies run: | @@ -53,10 +53,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Build - uses: cross-platform-actions/action@v0.15.0 + uses: cross-platform-actions/action@v0.24.0 with: operating_system: freebsd version: '13.2' @@ -77,10 +77,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Set up QEMU - uses: docker/setup-qemu-action@v2 + uses: docker/setup-qemu-action@v3 - name: Build ${{ matrix.arch }} run: | @@ -105,10 +105,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Set up QEMU - uses: docker/setup-qemu-action@v2 + uses: docker/setup-qemu-action@v3 - name: Build ${{ matrix.arch }} run: | @@ -133,10 +133,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Set up QEMU - uses: docker/setup-qemu-action@v2 + uses: docker/setup-qemu-action@v3 - name: Build ${{ matrix.arch }} run: | @@ -165,7 +165,7 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: add oneAPI to apt shell: bash @@ -189,7 +189,7 @@ jobs: - name: Clone id: checkout - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Build id: cmake_build @@ -215,7 +215,7 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: add oneAPI to apt shell: bash @@ -239,7 +239,7 @@ jobs: - name: Clone id: checkout - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Build id: cmake_build @@ -262,7 +262,7 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Setup ${{ matrix.sys }} uses: msys2/setup-msys2@v2 @@ -328,10 +328,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Add msbuild to PATH - uses: microsoft/setup-msbuild@v1 + uses: microsoft/setup-msbuild@v2 - name: Fetch SDL2 and set SDL2_DIR if: matrix.sdl2 == 'ON' @@ -356,14 +356,14 @@ jobs: run: copy "$env:SDL2_DIR/../lib/${{ matrix.s2arc }}/SDL2.dll" build/bin/${{ matrix.build }} - name: Upload dll - uses: actions/upload-artifact@v3 + uses: actions/upload-artifact@v4 with: name: ${{ matrix.jnaPath }}_whisper.dll path: build/bin/${{ matrix.build }}/whisper.dll - name: Upload binaries if: matrix.sdl2 == 'ON' - uses: actions/upload-artifact@v1 + uses: actions/upload-artifact@v4 with: name: whisper-bin-${{ matrix.arch }} path: build/bin/${{ matrix.build }} @@ -392,10 +392,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Add msbuild to PATH - uses: microsoft/setup-msbuild@v1 + uses: microsoft/setup-msbuild@v2 - name: Fetch OpenBLAS if: matrix.blas == 'ON' @@ -453,7 +453,7 @@ jobs: - name: Upload binaries if: matrix.blas == 'ON' && matrix.sdl2 == 'ON' - uses: actions/upload-artifact@v1 + uses: actions/upload-artifact@v4 with: name: whisper-blas${{ matrix.clblast == 'ON' && '-clblast' || ''}}-bin-${{ matrix.arch }} path: build/bin/${{ matrix.build }} @@ -476,14 +476,14 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Add msbuild to PATH - uses: microsoft/setup-msbuild@v1 + uses: microsoft/setup-msbuild@v2 - name: Install CUDA Toolkit id: cuda-toolkit - uses: Jimver/cuda-toolkit@v0.2.11 + uses: Jimver/cuda-toolkit@v0.2.15 with: cuda: '${{ matrix.cuda-toolkit }}' @@ -519,7 +519,7 @@ jobs: - name: Upload binaries if: matrix.sdl2 == 'ON' - uses: actions/upload-artifact@v1 + uses: actions/upload-artifact@v4 with: name: whisper-cublas-${{ matrix.cuda-toolkit }}-bin-${{ matrix.arch }} path: build/bin/${{ matrix.build }} @@ -533,10 +533,10 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Setup emsdk - uses: mymindstorm/setup-emsdk@v12 + uses: mymindstorm/setup-emsdk@v14 - name: Verify run: emcc -v @@ -555,7 +555,7 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Configure run: | @@ -573,24 +573,24 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 with: path: whisper - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 with: repository: ggerganov/ggml path: ggml - name: Install Java - uses: actions/setup-java@v3 + uses: actions/setup-java@v4 with: distribution: zulu - java-version: 17 + java-version: 21 - name: Setup Android SDK - uses: android-actions/setup-android@v2 + uses: android-actions/setup-android@v3 - name: Build run: | @@ -608,20 +608,19 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: set up JDK 11 - uses: actions/setup-java@v3 + uses: actions/setup-java@v4 with: java-version: '11' distribution: 'temurin' cache: gradle - name: Setup Android SDK - uses: android-actions/setup-android@v2 + uses: android-actions/setup-android@v3 with: - api-level: 30 - build-tools-version: 30.0.3 + cmdline-tools-version: 9.0 - name: Build run: | @@ -633,15 +632,16 @@ jobs: needs: [ 'windows' ] runs-on: windows-latest steps: - - uses: actions/checkout@v3 + - uses: actions/checkout@v4 - name: Install Java - uses: actions/setup-java@v1 + uses: actions/setup-java@v4 with: - java-version: 17 + distribution: zulu + java-version: 20 - name: Download Windows lib - uses: actions/download-artifact@v3 + uses: actions/download-artifact@v4 with: name: win32-x86-64_whisper.dll path: bindings/java/build/generated/resources/main/win32-x86-64 @@ -654,7 +654,7 @@ jobs: ./gradlew build - name: Upload jar - uses: actions/upload-artifact@v3 + uses: actions/upload-artifact@v4 with: name: whispercpp.jar path: bindings/java/build/libs/whispercpp-*.jar @@ -676,7 +676,7 @@ jobs: steps: - name: Clone - uses: actions/checkout@v3 + uses: actions/checkout@v4 - name: Test quantize run: | From adee3f9c1faec890eb0c5f3f6f2f73597a8b3962 Mon Sep 17 00:00:00 2001 From: Pedro Probst Date: Mon, 20 May 2024 03:08:48 -0300 Subject: [PATCH 11/17] node : add flash_attn param (#2170) --- examples/addon.node/__test__/whisper.spec.js | 1 + examples/addon.node/addon.cpp | 4 ++++ examples/addon.node/index.js | 1 + 3 files changed, 6 insertions(+) diff --git a/examples/addon.node/__test__/whisper.spec.js b/examples/addon.node/__test__/whisper.spec.js index 9ba86b62985..1ee888a1e00 100644 --- a/examples/addon.node/__test__/whisper.spec.js +++ b/examples/addon.node/__test__/whisper.spec.js @@ -12,6 +12,7 @@ const whisperParamsMock = { model: path.join(__dirname, "../../../models/ggml-base.en.bin"), fname_inp: path.join(__dirname, "../../../samples/jfk.wav"), use_gpu: true, + flash_attn: false, no_prints: true, comma_in_time: false, translate: true, diff --git a/examples/addon.node/addon.cpp b/examples/addon.node/addon.cpp index 8125e5dda4c..53bf1abb5a3 100644 --- a/examples/addon.node/addon.cpp +++ b/examples/addon.node/addon.cpp @@ -39,6 +39,7 @@ struct whisper_params { bool no_timestamps = false; bool no_prints = false; bool use_gpu = true; + bool flash_attn = false; bool comma_in_time = true; std::string language = "en"; @@ -146,6 +147,7 @@ int run(whisper_params ¶ms, std::vector> &result) { struct whisper_context_params cparams = whisper_context_default_params(); cparams.use_gpu = params.use_gpu; + cparams.flash_attn = params.flash_attn; struct whisper_context * ctx = whisper_init_from_file_with_params(params.model.c_str(), cparams); if (ctx == nullptr) { @@ -326,6 +328,7 @@ Napi::Value whisper(const Napi::CallbackInfo& info) { std::string model = whisper_params.Get("model").As(); std::string input = whisper_params.Get("fname_inp").As(); bool use_gpu = whisper_params.Get("use_gpu").As(); + bool flash_attn = whisper_params.Get("flash_attn").As(); bool no_prints = whisper_params.Get("no_prints").As(); bool no_timestamps = whisper_params.Get("no_timestamps").As(); int32_t audio_ctx = whisper_params.Get("audio_ctx").As(); @@ -346,6 +349,7 @@ Napi::Value whisper(const Napi::CallbackInfo& info) { params.model = model; params.fname_inp.emplace_back(input); params.use_gpu = use_gpu; + params.flash_attn = flash_attn; params.no_prints = no_prints; params.no_timestamps = no_timestamps; params.audio_ctx = audio_ctx; diff --git a/examples/addon.node/index.js b/examples/addon.node/index.js index 09b33c54024..643ee756452 100644 --- a/examples/addon.node/index.js +++ b/examples/addon.node/index.js @@ -12,6 +12,7 @@ const whisperParams = { model: path.join(__dirname, "../../models/ggml-base.en.bin"), fname_inp: path.join(__dirname, "../../samples/jfk.wav"), use_gpu: true, + flash_attn: false, no_prints: true, comma_in_time: false, translate: true, From 1b51fdf170714dcdd8fb9cfd02dcee684aac6150 Mon Sep 17 00:00:00 2001 From: William Tambellini Date: Tue, 21 May 2024 08:31:41 -0700 Subject: [PATCH 12/17] examples : add support for decoding input with ffmpeg (Linux) (#2133) - search for ffmpeg libs/headers at cmake time - added ffmpeg-transcode.cpp into libcommon if ffmpeg on - hooked ffmpeg trancoding in common read_wav(...) - passed test: ./main -m ggml-base.en.bin -f samples/jfk.mp3 --- CMakeLists.txt | 24 +++ cmake/FindFFmpeg.cmake | 163 ++++++++++++++++ examples/CMakeLists.txt | 5 + examples/common.cpp | 18 +- examples/ffmpeg-transcode.cpp | 350 ++++++++++++++++++++++++++++++++++ examples/main/CMakeLists.txt | 2 +- samples/.gitignore | 3 + samples/jfk.mp3 | Bin 0 -> 76447 bytes tests/CMakeLists.txt | 11 ++ 9 files changed, 574 insertions(+), 2 deletions(-) create mode 100644 cmake/FindFFmpeg.cmake create mode 100644 examples/ffmpeg-transcode.cpp create mode 100644 samples/jfk.mp3 diff --git a/CMakeLists.txt b/CMakeLists.txt index 588aa61cd11..3eb12c10783 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -59,6 +59,10 @@ option(WHISPER_BUILD_EXAMPLES "whisper: build examples" ${WHISPER_STANDA option(WHISPER_SDL2 "whisper: support for libSDL2" OFF) +if (CMAKE_SYSTEM_NAME MATCHES "Linux") + option(WHISPER_FFMPEG "whisper: support building and linking with ffmpeg libs (avcodec, swresample, ...)" OFF) +endif() + option(WHISPER_NO_AVX "whisper: disable AVX" OFF) option(WHISPER_NO_AVX2 "whisper: disable AVX2" OFF) option(WHISPER_NO_AVX512 "whisper: disable AVX512" ON) @@ -125,6 +129,26 @@ else() set(CMAKE_CXX_STANDARD 11) endif() +if (WHISPER_FFMPEG) + # As of cmake 3.27, there is no official cmake support for FindFFmpeg. + # Consequnelty we added a FindFFmpeg.cmake script the cmake subfolder: + # whisper.cpp does not need the full ffmpeg libs, just AVFORMAT AVCODEC AVUTIL SWRESAMPLE + # libswresample performs highly optimized audio resampling, rematrixing and sample format conversion operations + # libavcodec provides a generic encoding/decoding framework and contains multiple decoders and encoders for audio, video and subtitle streams, and several bitstream filters. + # libavformat provides a generic framework for multiplexing and demultiplexing (muxing and demuxing) audio, video and subtitle streams. + find_package(FFmpeg REQUIRED) + if (NOT ${FFMPEG_FOUND}) + message(FATAL_ERROR "Cannot find ffmpeg libs/headers") + endif() + message(STATUS "Found ffmpeg libs: ${FFMPEG_LIBRARIES}") + message(STATUS "Found ffmpeg headers in: ${FFMPEG_INCLUDE_DIRS}") + message(STATUS "ffmpeg definitions: ${FFMPEG_DEFINITIONS}") + message(STATUS "Found avformat ${AVFORMAT_VERSION}") + include_directories(${FFMPEG_INCLUDE_DIRS}) + add_compile_definitions(WHISPER_FFMPEG) + set(WHISPER_EXTRA_LIBS ${WHISPER_EXTRA_LIBS} ${FFMPEG_LIBRARIES}) +endif() + # on APPLE if (APPLE) # include Accelerate framework diff --git a/cmake/FindFFmpeg.cmake b/cmake/FindFFmpeg.cmake new file mode 100644 index 00000000000..19dc751605e --- /dev/null +++ b/cmake/FindFFmpeg.cmake @@ -0,0 +1,163 @@ +# From +# https://github.com/snikulov/cmake-modules/blob/master/FindFFmpeg.cmake +# +# vim: ts=2 sw=2 +# - Try to find the required ffmpeg components(default: AVFORMAT, AVUTIL, AVCODEC) +# +# Once done this will define +# FFMPEG_FOUND - System has the all required components. +# FFMPEG_INCLUDE_DIRS - Include directory necessary for using the required components headers. +# FFMPEG_LIBRARIES - Link these to use the required ffmpeg components. +# FFMPEG_DEFINITIONS - Compiler switches required for using the required ffmpeg components. +# +# For each of the components it will additionally set. +# - AVCODEC +# - AVDEVICE +# - AVFORMAT +# - AVFILTER +# - AVUTIL +# - POSTPROC +# - SWSCALE +# the following variables will be defined +# _FOUND - System has +# _INCLUDE_DIRS - Include directory necessary for using the headers +# _LIBRARIES - Link these to use +# _DEFINITIONS - Compiler switches required for using +# _VERSION - The components version +# +# Copyright (c) 2006, Matthias Kretz, +# Copyright (c) 2008, Alexander Neundorf, +# Copyright (c) 2011, Michael Jansen, +# +# Redistribution and use is allowed according to the terms of the BSD license. +# For details see the accompanying COPYING-CMAKE-SCRIPTS file. + +include(FindPackageHandleStandardArgs) + +# The default components were taken from a survey over other FindFFMPEG.cmake files +if (NOT FFmpeg_FIND_COMPONENTS) + set(FFmpeg_FIND_COMPONENTS AVFORMAT AVCODEC AVUTIL SWRESAMPLE) +endif() + +# +### Macro: set_component_found +# +# Marks the given component as found if both *_LIBRARIES AND *_INCLUDE_DIRS is present. +# +macro(set_component_found _component ) + if (${_component}_LIBRARIES AND ${_component}_INCLUDE_DIRS) + message(DEBUG " - ${_component} found.") + set(${_component}_FOUND TRUE) + else () + message(DEBUG " - ${_component} not found.") + endif () +endmacro() + +# +### Macro: find_component +# +# Checks for the given component by invoking pkgconfig and then looking up the libraries and +# include directories. +# +macro(find_component _component _pkgconfig _library _header) + + if (NOT WIN32) + # use pkg-config to get the directories and then use these values + # in the FIND_PATH() and FIND_LIBRARY() calls + find_package(PkgConfig) + if (PKG_CONFIG_FOUND) + pkg_check_modules(PC_${_component} ${_pkgconfig}) + message(STATUS "Pkgconfig found: ${PC_${_component}_INCLUDEDIR}") + message(STATUS "Pkgconfig found: ${PC_${_component}_INCLUDE_DIRS}") + message(STATUS "${PC_${_component}_CFLAGS}") + endif () + endif (NOT WIN32) + + + find_path(${_component}_INCLUDE_DIRS ${_header} + HINTS + ${PC_${_component}_INCLUDEDIR} + ${PC_${_component}_INCLUDE_DIRS} + PATH_SUFFIXES + ffmpeg + ) + + # CMake's default is to search first for shared libraries and then for static libraries. + # Todo later: add option to prefer static libs over dynamic: + find_library(${_component}_LIBRARIES NAMES ${_library} lib${_library}.a + HINTS + ${PC_${_component}_LIBDIR} + ${PC_${_component}_LIBRARY_DIRS} + ) + + set(${_component}_DEFINITIONS ${PC_${_component}_CFLAGS_OTHER} CACHE STRING "The ${_component} CFLAGS.") + set(${_component}_VERSION ${PC_${_component}_VERSION} CACHE STRING "The ${_component} version number.") + + set_component_found(${_component}) + + mark_as_advanced( + ${_component}_INCLUDE_DIRS + ${_component}_LIBRARIES + ${_component}_DEFINITIONS + ${_component}_VERSION) + +endmacro() + + +# Check for cached results. If there are skip the costly part. +if (NOT FFMPEG_LIBRARIES) + + # Check for all possible component. + find_component(AVCODEC libavcodec avcodec libavcodec/avcodec.h) + find_component(AVFORMAT libavformat avformat libavformat/avformat.h) + find_component(AVDEVICE libavdevice avdevice libavdevice/avdevice.h) + #find_component(AVRESAMPLE libavresample avresample libavresample/avresample.h) # old name for swresample + find_component(AVUTIL libavutil avutil libavutil/avutil.h) + find_component(AVFILTER libavfilter avfilter libavfilter/avfilter.h) + find_component(SWSCALE libswscale swscale libswscale/swscale.h) + find_component(POSTPROC libpostproc postproc libpostproc/postprocess.h) + find_component(SWRESAMPLE libswresample swresample libswresample/swresample.h) + + # Check if the required components were found and add their stuff to the FFMPEG_* vars. + foreach (_component ${FFmpeg_FIND_COMPONENTS}) + if (${_component}_FOUND) + # message(STATUS "Required component ${_component} present.") + set(FFMPEG_LIBRARIES ${FFMPEG_LIBRARIES} ${${_component}_LIBRARIES}) + set(FFMPEG_DEFINITIONS ${FFMPEG_DEFINITIONS} ${${_component}_DEFINITIONS}) + list(APPEND FFMPEG_INCLUDE_DIRS ${${_component}_INCLUDE_DIRS}) + else () + # message(STATUS "Required component ${_component} missing.") + endif () + endforeach () + + # Build the include path with duplicates removed. + if (FFMPEG_INCLUDE_DIRS) + list(REMOVE_DUPLICATES FFMPEG_INCLUDE_DIRS) + endif () + + # cache the vars. + set(FFMPEG_INCLUDE_DIRS ${FFMPEG_INCLUDE_DIRS} CACHE STRING "The FFmpeg include directories." FORCE) + set(FFMPEG_LIBRARIES ${FFMPEG_LIBRARIES} CACHE STRING "The FFmpeg libraries." FORCE) + set(FFMPEG_DEFINITIONS ${FFMPEG_DEFINITIONS} CACHE STRING "The FFmpeg cflags." FORCE) + + mark_as_advanced(FFMPEG_INCLUDE_DIRS + FFMPEG_LIBRARIES + FFMPEG_DEFINITIONS) + +endif () + +# Now set the noncached _FOUND vars for the components. +# whisper.cpp does not need SWSCALE +foreach (_component AVCODEC AVDEVICE AVFORMAT AVRESAMPLE AVUTIL POSTPROCESS) + set_component_found(${_component}) +endforeach () + +# Compile the list of required vars +set(_FFmpeg_REQUIRED_VARS FFMPEG_LIBRARIES FFMPEG_INCLUDE_DIRS) +foreach (_component ${FFmpeg_FIND_COMPONENTS}) + list(APPEND _FFmpeg_REQUIRED_VARS ${_component}_LIBRARIES ${_component}_INCLUDE_DIRS) +endforeach () + +# Give a nice error message if some of the required vars are missing. +find_package_handle_standard_args(FFmpeg DEFAULT_MSG ${_FFmpeg_REQUIRED_VARS}) + diff --git a/examples/CMakeLists.txt b/examples/CMakeLists.txt index 3b493e3db7e..24678e1c6ac 100644 --- a/examples/CMakeLists.txt +++ b/examples/CMakeLists.txt @@ -22,6 +22,10 @@ endif() set(TARGET common) +if (WHISPER_FFMPEG) + set(COMMON_SOURCES_FFMPEG ffmpeg-transcode.cpp) +endif() + add_library(${TARGET} STATIC common.h common.cpp @@ -29,6 +33,7 @@ add_library(${TARGET} STATIC common-ggml.cpp grammar-parser.h grammar-parser.cpp + ${COMMON_SOURCES_FFMPEG} ) include(DefaultTargetOptions) diff --git a/examples/common.cpp b/examples/common.cpp index 2c0cdf082ed..25a0272cf08 100644 --- a/examples/common.cpp +++ b/examples/common.cpp @@ -24,6 +24,11 @@ #include #endif +#ifdef WHISPER_FFMPEG +// as implemented in ffmpeg_trancode.cpp only embedded in common lib if whisper built with ffmpeg support +extern bool ffmpeg_decode_audio(const std::string & ifname, std::vector & wav_data); +#endif + // Function to check if the next argument exists std::string get_next_arg(int& i, int argc, char** argv, const std::string& flag, gpt_params& params) { if (i + 1 < argc && argv[i + 1][0] != '-') { @@ -637,7 +642,7 @@ bool is_wav_buffer(const std::string buf) { bool read_wav(const std::string & fname, std::vector& pcmf32, std::vector>& pcmf32s, bool stereo) { drwav wav; - std::vector wav_data; // used for pipe input from stdin + std::vector wav_data; // used for pipe input from stdin or ffmpeg decoding output if (fname == "-") { { @@ -670,8 +675,19 @@ bool read_wav(const std::string & fname, std::vector& pcmf32, std::vector } } else if (drwav_init_file(&wav, fname.c_str(), nullptr) == false) { +#if defined(WHISPER_FFMPEG) + if (ffmpeg_decode_audio(fname, wav_data) != 0) { + fprintf(stderr, "error: failed to ffmpeg decode '%s' \n", fname.c_str()); + return false; + } + if (drwav_init_memory(&wav, wav_data.data(), wav_data.size(), nullptr) == false) { + fprintf(stderr, "error: failed to read wav data as wav \n"); + return false; + } +#else fprintf(stderr, "error: failed to open '%s' as WAV file\n", fname.c_str()); return false; +#endif } if (wav.channels != 1 && wav.channels != 2) { diff --git a/examples/ffmpeg-transcode.cpp b/examples/ffmpeg-transcode.cpp new file mode 100644 index 00000000000..910cdf5700b --- /dev/null +++ b/examples/ffmpeg-transcode.cpp @@ -0,0 +1,350 @@ +/* SPDX-License-Identifier: GPL-2.0 */ + +/* + * transcode.c - convert audio file to WAVE + * + * Copyright (C) 2019 Andrew Clayton + * Copyright (C) 2024 William Tambellini + */ + +// Just for conveninent C++ API +#include +#include + +// C +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +extern "C" { +#include +#include +#include +#include +} + +typedef uint64_t u64; +typedef int64_t s64; +typedef uint32_t u32; +typedef int32_t s32; +typedef uint16_t u16; +typedef int16_t s16; +typedef uint8_t u8; +typedef int8_t s8; + +#define WAVE_SAMPLE_RATE 16000 +#define AVIO_CTX_BUF_SZ 4096 + +static const char* ffmpegLog = getenv("FFMPEG_LOG"); +// Todo: add __FILE__ __LINE__ +#define LOG(...) \ + do { if (ffmpegLog) fprintf(stderr, __VA_ARGS__); } while(0) // C99 + +/* + * WAVE file header based on definition from + * https://gist.github.com/Jon-Schneider/8b7c53d27a7a13346a643dac9c19d34f + * + * We must ensure this structure doesn't have any holes or + * padding so we can just map it straight to the WAVE data. + */ +struct wave_hdr { + /* RIFF Header: "RIFF" */ + char riff_header[4]; + /* size of audio data + sizeof(struct wave_hdr) - 8 */ + int wav_size; + /* "WAVE" */ + char wav_header[4]; + + /* Format Header */ + /* "fmt " (includes trailing space) */ + char fmt_header[4]; + /* Should be 16 for PCM */ + int fmt_chunk_size; + /* Should be 1 for PCM. 3 for IEEE Float */ + s16 audio_format; + s16 num_channels; + int sample_rate; + /* + * Number of bytes per second + * sample_rate * num_channels * bit_depth/8 + */ + int byte_rate; + /* num_channels * bytes per sample */ + s16 sample_alignment; + /* bits per sample */ + s16 bit_depth; + + /* Data Header */ + /* "data" */ + char data_header[4]; + /* + * size of audio + * number of samples * num_channels * bit_depth/8 + */ + int data_bytes; +} __attribute__((__packed__)); + +struct audio_buffer { + u8 *ptr; + int size; /* size left in the buffer */ +}; + +static void set_wave_hdr(wave_hdr& wh, size_t size) { + memcpy(&wh.riff_header, "RIFF", 4); + wh.wav_size = size + sizeof(struct wave_hdr) - 8; + memcpy(&wh.wav_header, "WAVE", 4); + memcpy(&wh.fmt_header, "fmt ", 4); + wh.fmt_chunk_size = 16; + wh.audio_format = 1; + wh.num_channels = 1; + wh.sample_rate = WAVE_SAMPLE_RATE; + wh.sample_alignment = 2; + wh.bit_depth = 16; + wh.byte_rate = wh.sample_rate * wh.sample_alignment; + memcpy(&wh.data_header, "data", 4); + wh.data_bytes = size; +} + +static void write_wave_hdr(int fd, size_t size) { + struct wave_hdr wh; + set_wave_hdr(wh, size); + write(fd, &wh, sizeof(struct wave_hdr)); +} + +static int map_file(int fd, u8 **ptr, size_t *size) +{ + struct stat sb; + + fstat(fd, &sb); + *size = sb.st_size; + + *ptr = (u8*)mmap(NULL, *size, PROT_READ|PROT_WRITE, MAP_PRIVATE, fd, 0); + if (*ptr == MAP_FAILED) { + perror("mmap"); + return -1; + } + + return 0; +} + +static int read_packet(void *opaque, u8 *buf, int buf_size) +{ + struct audio_buffer *audio_buf = (audio_buffer*)opaque; + + buf_size = FFMIN(buf_size, audio_buf->size); + + /* copy internal buffer data to buf */ + memcpy(buf, audio_buf->ptr, buf_size); + audio_buf->ptr += buf_size; + audio_buf->size -= buf_size; + + return buf_size; +} + +static void convert_frame(struct SwrContext *swr, AVCodecContext *codec, + AVFrame *frame, s16 **data, int *size, bool flush) +{ + int nr_samples; + s64 delay; + u8 *buffer; + + delay = swr_get_delay(swr, codec->sample_rate); + nr_samples = av_rescale_rnd(delay + frame->nb_samples, + WAVE_SAMPLE_RATE, codec->sample_rate, + AV_ROUND_UP); + av_samples_alloc(&buffer, NULL, 1, nr_samples, AV_SAMPLE_FMT_S16, 0); + + /* + * !flush is used to check if we are flushing any remaining + * conversion buffers... + */ + nr_samples = swr_convert(swr, &buffer, nr_samples, + !flush ? (const u8 **)frame->data : NULL, + !flush ? frame->nb_samples : 0); + + *data = (s16*)realloc(*data, (*size + nr_samples) * sizeof(s16)); + memcpy(*data + *size, buffer, nr_samples * sizeof(s16)); + *size += nr_samples; + av_freep(&buffer); +} + +static bool is_audio_stream(const AVStream *stream) +{ + if (stream->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) + return true; + + return false; +} + +// Return non zero on error, 0 on success +// audio_buffer: input memory +// data: decoded output audio data (wav file) +// size: size of output data +static int decode_audio(struct audio_buffer *audio_buf, s16 **data, int *size) +{ + LOG("decode_audio: input size: %d\n", audio_buf->size); + AVFormatContext *fmt_ctx; + AVIOContext *avio_ctx; + AVStream *stream; + AVCodecContext *codec; + AVPacket packet; + AVFrame *frame; + struct SwrContext *swr; + u8 *avio_ctx_buffer; + unsigned int i; + int stream_index = -1; + int err; + const size_t errbuffsize = 1024; + char errbuff[errbuffsize]; + + av_register_all(); // from avformat. Still a must-have call for ffmpeg v3! (can be skipped for later versions) + + fmt_ctx = avformat_alloc_context(); + avio_ctx_buffer = (u8*)av_malloc(AVIO_CTX_BUF_SZ); + LOG("Creating an avio context: AVIO_CTX_BUF_SZ=%d\n", AVIO_CTX_BUF_SZ); + avio_ctx = avio_alloc_context(avio_ctx_buffer, AVIO_CTX_BUF_SZ, 0, audio_buf, &read_packet, NULL, NULL); + fmt_ctx->pb = avio_ctx; + + // open the input stream and read header + err = avformat_open_input(&fmt_ctx, NULL, NULL, NULL); + if (err) { + LOG("Could not read audio buffer: %d: %s\n", err, av_make_error_string(errbuff, errbuffsize, err)); + return err; + } + + err = avformat_find_stream_info(fmt_ctx, NULL); + if (err < 0) { + LOG("Could not retrieve stream info from audio buffer: %d\n", err); + return err; + } + + for (i = 0; i < fmt_ctx->nb_streams; i++) { + if (is_audio_stream(fmt_ctx->streams[i])) { + stream_index = i; + break; + } + } + + if (stream_index == -1) { + LOG("Could not retrieve audio stream from buffer\n"); + return -1; + } + + stream = fmt_ctx->streams[stream_index]; + codec = avcodec_alloc_context3( + avcodec_find_decoder(stream->codecpar->codec_id)); + avcodec_parameters_to_context(codec, stream->codecpar); + err = avcodec_open2(codec, avcodec_find_decoder(codec->codec_id), + NULL); + if (err) { + LOG("Failed to open decoder for stream #%d in audio buffer\n", stream_index); + return err; + } + + /* prepare resampler */ + swr = swr_alloc(); + + av_opt_set_int(swr, "in_channel_count", codec->channels, 0); + av_opt_set_int(swr, "out_channel_count", 1, 0); + av_opt_set_int(swr, "in_channel_layout", codec->channel_layout, 0); + av_opt_set_int(swr, "out_channel_layout", AV_CH_LAYOUT_MONO, 0); + av_opt_set_int(swr, "in_sample_rate", codec->sample_rate, 0); + av_opt_set_int(swr, "out_sample_rate", WAVE_SAMPLE_RATE, 0); + av_opt_set_sample_fmt(swr, "in_sample_fmt", codec->sample_fmt, 0); + av_opt_set_sample_fmt(swr, "out_sample_fmt", AV_SAMPLE_FMT_S16, 0); + + swr_init(swr); + if (!swr_is_initialized(swr)) { + LOG("Resampler has not been properly initialized\n"); + return -1; + } + + av_init_packet(&packet); + frame = av_frame_alloc(); + if (!frame) { + LOG("Error allocating the frame\n"); + return -1; + } + + /* iterate through frames */ + *data = NULL; + *size = 0; + while (av_read_frame(fmt_ctx, &packet) >= 0) { + avcodec_send_packet(codec, &packet); + + err = avcodec_receive_frame(codec, frame); + if (err == AVERROR(EAGAIN)) + continue; + + convert_frame(swr, codec, frame, data, size, false); + } + /* Flush any remaining conversion buffers... */ + convert_frame(swr, codec, frame, data, size, true); + + av_frame_free(&frame); + swr_free(&swr); + //avio_context_free(); // todo? + avcodec_close(codec); + avformat_close_input(&fmt_ctx); + avformat_free_context(fmt_ctx); + + if (avio_ctx) { + av_freep(&avio_ctx->buffer); + av_freep(&avio_ctx); + } + + return 0; +} + +// in mem decoding/conversion/resampling: +// ifname: input file path +// owav_data: in mem wav file. Can be forwarded as it to whisper/drwav +// return 0 on success +int ffmpeg_decode_audio(const std::string &ifname, std::vector& owav_data) { + LOG("ffmpeg_decode_audio: %s\n", ifname.c_str()); + int ifd = open(ifname.c_str(), O_RDONLY); + if (ifd == -1) { + fprintf(stderr, "Couldn't open input file %s\n", ifname.c_str()); + return -1; + } + u8 *ibuf = NULL; + size_t ibuf_size; + int err = map_file(ifd, &ibuf, &ibuf_size); + if (err) { + LOG("Couldn't map input file %s\n", ifname.c_str()); + return err; + } + LOG("Mapped input file: %x size: %d\n", ibuf, ibuf_size); + struct audio_buffer inaudio_buf; + inaudio_buf.ptr = ibuf; + inaudio_buf.size = ibuf_size; + + s16 *odata=NULL; + int osize=0; + + err = decode_audio(&inaudio_buf, &odata, &osize); + LOG("decode_audio returned %d \n", err); + if (err != 0) { + LOG("decode_audio failed\n"); + return err; + } + LOG("decode_audio output size: %d\n", osize); + + wave_hdr wh; + const size_t outdatasize = osize * sizeof(s16); + set_wave_hdr(wh, outdatasize); + owav_data.resize(sizeof(wave_hdr) + outdatasize); + // header: + memcpy(owav_data.data(), &wh, sizeof(wave_hdr)); + // the data: + memcpy(owav_data.data() + sizeof(wave_hdr), odata, osize* sizeof(s16)); + + return 0; +} diff --git a/examples/main/CMakeLists.txt b/examples/main/CMakeLists.txt index 1bb16f58214..1e66e4b5cc8 100644 --- a/examples/main/CMakeLists.txt +++ b/examples/main/CMakeLists.txt @@ -3,4 +3,4 @@ add_executable(${TARGET} main.cpp) include(DefaultTargetOptions) -target_link_libraries(${TARGET} PRIVATE common whisper ${CMAKE_THREAD_LIBS_INIT}) +target_link_libraries(${TARGET} PRIVATE common whisper ${FFMPEG_LIBRARIES} ${CMAKE_THREAD_LIBS_INIT}) diff --git a/samples/.gitignore b/samples/.gitignore index 72e8ffc0db8..e084659df25 100644 --- a/samples/.gitignore +++ b/samples/.gitignore @@ -1 +1,4 @@ * +!jfk.wave +!jfk.mp3 + diff --git a/samples/jfk.mp3 b/samples/jfk.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..fbfa1d9897365f05adf4f624c3210cc4fd374a44 GIT binary patch literal 76447 zcmce7_g7P2@a;_{K!8vUND0-@L$6BcT|$+rp%)P;N>P;1yYwm|ND&Z}-b4vix^zXU z7EnY$X&(^CgYSE9t@jtav(`=Oe3fMfxXhz9_H zQqj^gv7$M+FnmJ7q7pb6IRzz^%U9Gjv~=_gOf0V1+B*>4Ja69e@%6tG7<@036cHVl zkerg9nfth~xTL)D#mn0IhNhO*_RijSgCk=TQ!{g)7ni?&|FOQg{daf&zoV10|05Ww z8*9s)TLUHhA4>pMga8E33xJl-?cjqx>iWMo{y)D{c7l5WSTGn51_1B{%Ktl_-o7M! z^#1q?5O%`1*K$rT z{5MG{1bZYsnzDoKons88*W$sWKqyUVT2dly7=@yC$eP&yvn<=QNEa2?D+Wd5hEod- z;Cf;hYVr45;sJRSg;#Opjkkhe>8q3Ylw$|VPHyj*StgsZL%~PjA4!9(R~<`2jaPV( z!DUP!3C#@=*a_v75>lu-ZYD-`~RSY`~iOi(;EM9Kb>gUu7b#R4L z=XGcE=|3YL6~wqw7`BdK_?gmlw~d0=Q)CHikXd0Lk4(3)(*;nEEDmbMko8{A^>6Mz zq4%mCm7!;rKn_Qo{CtC*nSzA7%O^tZaIaY!`ipT`xnil^om~fujkwM3LO0_}QU+x2YDmRulc0`S8>7ub1ji zKfe)tWg^s|roh!Cm9ZlJjI<`$y4Yr@q#MP9`u(H8R#HW{#rUC=iEWIHWT$%8#n&Qj zUJGI~wsucH2MHGs^tQjOD_@&`7o?JNb1nPwP=EFJ>>D`(4cD&xerwSg{p)j`&)CLL z?bytS*lmjB*3QKPx#MBFfTmkP5{svf2j{W6s9Lfgc;h1(L}^mF-x8uE~D*f>0pIT6n zg*sd5RZu4EP9LJjgY47SoeOy@~?+(_8@Peu2c#0+@P*UX}m!E*0TXuZP5rc z4;RhoOn4g~ze}9)-)MO+6_}AKfr^i{ddXs@GSxLPIHXKNM|#pjKljnzVIZ?()O62F z_3-fFClv?rWVPCnFE!lhgpAadzbXDzo3`75;wapSLzYNBIkJQT=_|>ffX1HZNl>=%brM zVl7`Q{k%9Y_vS(g+DwJTYb5Q>gMVLj-d^tftc;0I&Z^nsRvq=<-`!Z4UpkOi)n1N1 z`1^Kz^n1o(S_p36`r#h(L3~mh-B8``Pm|K4lT#hvx;ch_LLRJna>4?P^)%3@$iSV* zs<7Wnm0Kl^W~@CG$EIycJo@492` zqF1l|H^53o(7qf)&wO|(()UM{EO9Ns$~p%3H-_f&7tU(l+z%}p=QjVsM*fZ44~b3g zGvM~mzw@R4_ScCa*|c|l!i?qto{;oHm9%~Ltxp1a(`~^wn?wDD>L>ox?q$q}l-!Jw zdr2qB-9%wNE;G6Fw0HQKg~Q#Zi!CF*-_4EhYkWFkb>HT(6ag61B5bf<+TJcHl@FwU zk>w%VIZ=T$V#;q9SXB`GhlSMRuz!jj318|Mr7%pa^mNpqA#{ocV;M(xQ)PKa*Yu!# z?vQ%5o=D))sh`&i%8p*|HkNiX>VpX$Wp!k>PhbsZc$qA_*}85)$g+vrcVB*PMtg95fQ;n{aG5)m&+1>#(8c&lEcKm-cGs z96o7k&*qtvM+T?^ey%VLu+Qc{D5CnSanxC}krKR~M_Y7Z>}WQ+yNv$vfk?*Zp^dtO z+u5o-Jbqs$gnafj8>asoCaPTfaibzom-YFRoQ3XdbPgq*kl7a|>KrZbnX=9xs=39n z!Y7{_3^jfFU-QRsB>(h&>$9rbAhlnbg+KrR6;~P`EE7x;N))N>M~}FMmt#;N@j?O? zK|r*NMz>MYwDIU8i}8y|(9A*Jf!<=-o^i4F4O&xo3O=bTj%o~iPTTFf^rL&6r5le6 zhTyEx(l{rPaQG$aZgTj?XusTW#{|cpIpL;}>S!zv4g

P~0>~dJsUCB3wX4Mgyp* zI6feuXfCo#J(_fw2m*zd+Qfmn(wA3?fvyPNJS?IKY0Wg!%$cMNN=BO@^(B)=IFeX2 z>EL=vR~fHbBaDiWMuPO+;RqZ*4lOQT8SShX)gd*VC1vDKxA|u&!G+chfu?W!g!T34!GA<2FUFtd$j6GDq3P!0M z&syRa)Oja}OkN2mn-aKinbgIMfL>oV`ufkPm5DU4QY5*S__x{#C9LoKBILN-EuA_^ z>!!o+ME}ZLdXMaCKH0xcIoJ!LKa&S%<5YI)cB#ARYhUL|#u!nZmz{Sp`YvxD(;U6J z_%y-gSo*NtYudbV_V-ueOvl6nuLQSyg%J$#^UoV(tQc)?xH6O(v-yV?rFr8q>Iq-s z^RyV_fiGM=AiSQU04P_7n;kBnw_U&y?1~P?NHeo&X`%+dURy3nTb#1Hd!pxcb4h9-U6~8L8C*SC(Y6hH7qbH0 zh!}4o5${-G;Rv)7$9tW8iOwN#Cv>1QmXoNbO8mp^yiM;q80Cn$>K3l&5})9qbd|Ex z+FR{q0k@}G_yHoCJ-h8I;GBO9br^Z?mn;&V!Y8!XPvlH^94;HDG4icHZg2CyPh_7k}2 z7gDN@I8IXtmkxcB$zik>C{7bG!N*D0^U6y270Pwb%iQ+3cCqt+=I#9D2Bq`E7A5Re z52f>n((`obz1r!i;@Onmc>j*_ujBP$(B|;J!?0hJ(#pQ44^~J^ z;Iw5B6cQ{8G6ItXT#!aQU`&HS0a+EPPL?Jrf?Y`lI3CzndedJxVNeEy#EcUMiiqG` z@o>{9Tuz=KO-OtMZ?1HbFP4W)@*o-!bA^bAdZ&-yT=bxxgMnyS8bG4WS5KXvt{=jz z?F?u5$^cmGh4Bm<&lE)JM&cs=6md9uBGf@7V$U+Zm-yp72#@=UWauV!MLXgIe54E3 zWgLU$5|5=x6kKB)00;@Pm)}D6o3pu@pv#4n$|GoEF2t^KfWJl5#dk|RBDL0Dr6-Pz zS|^!F1Wf88_;wc^ z_oRxEfeLbjSU<0gIv&r$ddV_+7)QY3yq$5a3=S^jZ@(h}5@;FKC8gL`)W#y+MEZen z{K4tSIPHWoqY;b`7g#0bBcL~=%#xH6WN=~tkR*hI>)6NYDHu0hJ@m|^9WhFeRNS62i7ZH3 zK2cHE7h<$_AyUWNF0}YL8VR#Nt-fTzM|{1B*mv4lo)AZU(^ME+ZjbuN4_3ug7KajX zFK!VW0qFTtw3)sZ?)1e#5+O%wm(7Fz`H?tjSeNeTUfmucdSLIDK7h#Z(3l_zWK%!ks*Z4vy4L$i3m zdqJcOs{BzG8!E72`;;3`v&M~Wz#4$R@`S$cFKsh z3!j=kCP38{&=6>l)91}P?{k8n4%83NA>byIpwsh zd3artyU1QbS`>zkQw*7IvH3O~;r_8d-N&mj+04q9e3*dEHv--E+)7m7B{qoxkMb@L zW-{;!K=E;U7B;?!!|Z4#9`wMMH%`cW4lmeGI4iZO>hK0mj2xb&+?)$sOuJgfU=b6t z{fbUIFH$X|K0AQ(zLO=pk%~ynNevy$r16>`Y502hI2GuL^!He!nE<09$B#WNg4xnJ zhiyhL=XvvJw=hpUe=(|885UeHlm0u@?R@Y0Z+2zA#NJ@zw| z${XLh`AgmWzA%@e^*rkJ2g`em3zKh|*(&P=%HFK;s-5xbf3H4>@D#w$YqV)4$lGA} zXU4v0pSOqY#_vjFI&-Ktw0p{dJF@g*LT9CRUU~%mQ{+KzH}9HgZY!{}5)UTjAZyTy zaT8gisvCnqQmPoH66B_L-RYSdJ&lXkySLs>s~RkD1lG}V)+r*FNs>KN=sj&jSyipZ zW#t&-7ax^0X4W%|YR?pS9Cs@vbgsURoI7eI#7oiX1`g@~X3r)o&kKd|{S-10q!68~)cEt#f3b zmgsaxNv}3G&qM#AP!#stwH)ci=)9QVhqPv0!MYdUpL6Dt<^6IC_+sJVSj2vupJrBk zpoAS`{NIn^vfve3jbTTqd<=V-4QF(;R>7ri4`UkB%Mc}g?UZNZot6EmO$C3}c}U}a zm%_=Qy#99W@~IUDW;xdD6`%9@rw%&|$uT`}*0;|0Z(F~BDiwE_)K<52JCz9N+;|HW z(cKRXC^bI-?sPI5V6Ov*Lf=7Q&CQ=Q zzK8Hrb97n%f*ZR>YuKj-)s9BC7j{L$Kl=k1f)be`?Q&`BrO55|MDe2EP$3XsqG0bPiKBix#}f-?$8YXd>(Y z5{w2TnEx6Iz=}akrY)LOi}ntU3C402&*ELay)5$l)W8>eR5 zjLEhQCuX#FMg{!bUOvA6E;1~L@~i(%*yk+0EFf^w@jos>Fk6FTFID>w3{gCJ&vDul z+OhatJ+RH&p*5S64!FP2T5v>t@$uq!yy*}&(lJb0@}ga$9`W5h)aeBoJ-ew`^R&!3 z^^1s?M82_|G?&O|#I7a+bz3x1=biJQ?e@bAg|d5-54h94B;4l0B&&#&RM!6tFjXfpN#bl2G5A5y`GK4DCw#qx13O?HBQmS}&M2 zN$ucs|1{xy+P&dD@9at6@jMB8W$0HqyF!}vniik0Ut?pz7`&NtxydGtj!Ab~=J^Ut zO2qsePp%8ST6=GXZ@K&5y)WPLL=Glu-rTq=9R1V!>BM(M&aUUG((FK(4BJDS|A?0` zwU>!d5<577#$_P;B&Ax!u(2YKx*?*skDwR}!jPkq(xP2%eUZ+2ki_Sd-$^Ic6|#5` za-IKZE8lT@xOGkH+Ab5_jYnWjW;Pg>&2IL^ED^0~C~?t@xFIFO`BU?><6Hf~pOwd@ z4+SUL%`MFDnux_Kz{uOuI*QL7eKC%nDMc9F@YEI^%=m*wb|0GTkM19@%e1le@z2Q4 zSk=CJagVKq##+hs-jQ;c&Q1iYs4|C1{$tjwlfwdMeZRQgJ}p`LVEEL`=u+@KiLs1I zJ>_(0$GpZ0=clf+sDGus8wnPQizZK?b2h3}TY0W`8rwho`DMGt#h}S>xxC@m($d1+ zhA|Nc;%Y@or68)EyUKAuEei-Osf)91$bdS@X0t?OI&ie@!fPx=Y_4 z+K7wuvL~9o0IKJ?_l*s2uv)FB)+2U3Cf}xb9osRoZsLZ~jGx zLX&}VcCb3>BZgKKg%pC0C|X>tk3qJwM#6~|^h5wfpb;#>sNqnO(u)7&IAE%&%To6I z#Bg!~brKASAqknWQAdD%J&45id(Iw{aOx@JS5>}kysq(Zq$+X106E>*10{hl%P|gC z9D@`%EHdUshBTH9W18piT<`=Hhi=^#9_R-#Bzvzg!C-i@Yme8a*<Rh|8_42}n#Khj->~Ujbn|#tLr2xWqTp}j%7kZCulQe?a|iI8y26(pKlwd$>)@_2Ga~xafHJE< zh}FdALHt{%hQl?F7!eT;>lwN4cyvj5KIe71lB-p5BEW;pD*B-5e5PK~7cwBmOa<+y z2`72XMMLl<0J${0`fD!0LQU<&NFWRJZkuR*Y3S`DTBkz9{(@1^aU%vu5(6la?5z0m z_MZWiNDhKrq=%t#A2@vi2S=`Qz~E}-I(Gzhl%QRpZb7fSUrxfM-AF0Una9X)0zF+9 zJ4~xvJRy6sP&%2?i8Q>F=3vnFf7%vdg0vv9JnUnenA$PVnx^qq&lKhSPw|sB7YX%{ z%KGx>r0W>GO&)bF8k7K5X$bTzN#QV{k7K_l6`n)VxbXdaz0VjoygiHO_cn6k&_m&0=qbK~tJ z@bW-h(=D~r+kbX>-i}(?iH}}p=~Xhpt;+sht-UflY{rwz&F)kPkfCRN=K|$&rfb^J zDsLki2%yOg3nz);Xuu@F|JXpfBr?kXNeq}&E-a+;$2V3!GZOH|)&8DkqKs5W!eSxH z*_sBlrbUj6I3$y==TDEx86G4xGARWhYp#16-n2_MSaK!EP}A*eF2>*QEJ#w3kwQm) z!k};FM&wR@r_Ay7y8rw^L~@wUy39(KHC8fnc2Ox?f}2N+6arsb3V$CXFOUswa%|yw|cvHz&N>*?!o3RYZW0uAUsokPE&2Px;cDOvwW(7wJ#DDo<`)7 z`#KTNqW&@$s>1<)mTM1la&XZSqWX6Hb{;}P8dMKzL84tR)|i&ncz9kxmT}^;WQy07 z3=5B!`sQ#_wFG3yQD^4|m(*>j_W;P!Y4-Jt8*fEYi#-R*n9-#0lB-P$qVKP)5Y<=i ztcQ68<%1 zzvtaodx&mkBW{6PPj?EoXjznffAXZOo#^vWG?vahkn&mn#oeI`&)#scwArvCU|}O2 zsrv3y62TS-4_bO6dfur*BEqdBGP?C@#P0|^4ngCfDQsc>$&~1;L)^v==mLM1dAkS@ zFxeNtI0h@i~!)8O12hwur%54Y8>E94aJi!H=p+e%eDrJ{LZE7AlOE(SjLU5 zn~{4x>#&0VdTgfanga`}pWfoTFTQ`G{m5hT{zv;6F?>sVZzonL6dICk{wjLwwfa4? zg~VDvgZPZR4>O&lPsIcfO}sAdpQ-z8ipAtwU^B~p^^HCD2)=$}FSTF00%m;X*VdR) zI{UtyYyMBFNlc;To0b-XqbEb$51m6OloP%6@(kOmYNtS{joGv+64gT+{G*HfR)0vW z79#+Fz-us}`Uz6Raf2L35{N+MwL&rtMe-^c2P}>tSR%re!Z9n6%SwUyqu2@r9)z}1 z&f{~-7n6?#y;AiNl*}AR0)a;8=!|rMNHE-(W0oa7#;otcPZG&pe6@*f)o;JiMqW%c zO3?bS;P-mI4iOY&c=xjL>-09e$>CWvnn~czEx{3Mx7+on016Nhmav$(U-45|%&I+q zJ$2~&JlS3NZ%hqLhekP6l%p@)XI&EEL+9lGT&jFV*6VoptptD+HvD!T8?$OIY-iYE z77uP%tRkF2kNh(kB9+4C$YK3Zs5ip|g>poq-zQ3Oc_-wQI}y-TB=uYWDlL?Y<-8RF zM;rlOQvc8eI@}(QI*pIPAV>)9#hKrI$AaKAPXbcDk_V3GIk9jYCU{V5hZdv>xPE)% zquz32sU2QaBqQrBx4Pz7{PRx?lVgVywxFhQa2`6q3#ikvfNAQE>fbNY{QOex?w5jU zskdotFI1ZMXjT~j)Kg0pqm?vg*MEO)Rsn5`vN3U0Ln5_C7F!xQT ziX2##yJ1KnyF9HWe>WU$46p{Vy0sI6Dq(euS7PVByk*T8i9PB}()_bSS<+o9TQ6h? zdm`KJ88C0D*qw|1p-9>B==~+X^ERsg>#*gd$Q8}89oTawxC=Id^{V`RNW3f%t0J1s z3~qt;&_w29|KvntS~oytrUJPD%8|Lkm;XtIAsQ3q;mkf99p~lKQDZO{zdqX1nt&Cj zFuuu+k9YESatr#!8s4O{SjQ0}v*ug2dn8}?ST{cID0@FqR0RovA|%3p&T!rLo>Zn4 z<&mstqc=C&mC1=0)wXt`{8d|$c9{89F*ZO+dElby7Bn}dYruJ=*h?E0rFJ3xx%78^ z?YIm$A3pCtVg%b+B^Wc|nHrgNrja~)KS}hm#lb%hqfxlDkyXhl#Z1sZW-?9@wo_us z426!^y2%g_z6d@@&Tdov&0%37BP$i@!QDZOHOG7wuipcv5hxW7{iIvNuI1jzdh|Bi z%3Gwa$Yc^S3Lyf;aG;jWX?mA4pmCkWw(pk&br`Eb82{YqvFSBi4&FG;*jXYDiz?Zc z>-q0G|Ly3xXW2X&Xkgp!fI@os`gPgstt%lluV#C>gBnt5=qRH~-Q9ZG_bLBQ{A;ib z5z=dBiN=F+YYOuv7NK7!Qi9g??#^74xI%l5Yh78Eh6Mlm-jC95?>^a$s=Ym4fh9*Q z`|TY_FmLDnkOm`cXsAX$E6DOj4Mdv(ww_IlM9w-Q9`5`&AnBeSF4#Iw!75;74G@!% z^E3MJ@~ceb)5Bl>2HLkxO9W_BT_ckr?X<`0A{d7Its^pk9#_A{cOMPDq5-LNEoM*rIFlg8 zh`2~5hN0Ed3*bcw*-LmNAGWFerAsVd3ZcuM!XKq=!fudN}wEx zgzf9P;-op6zc`e|z_T5eHhljw!7pA(Hr*XmNYdiaEGEBPoC#fQ9F}rQZo2?dcb#<& z{PyF&a)Iurr)Fza(|oAXgdtk3m9{4)S2f7p$pX$s*xbupPv#p30Fk7-IKLkc^Vj^0 z-+}b=72N=7Lg!L;^Ir9ybK%;_IY*7E8SdYqU(NCxo6o%ND7h4t8jjb)s2s0RD5UMJ zvt&inJ-F1Aac{cd=)KgZiU+2aR&n*_EN%}i!M7#hJe;kl zEgW#fGwmmVipi*o0at#43%Y_>tK&$f5kn@<-)D_Lf z`tr>Il6qGRG%Z{cMHJgQc2hjnOTzzW&4AooLzGLhSXt}fK zh(I%?ek_KJzwIf)Ry*_cLB{0EedEjU1Z3AYBeFdw1gTekpU2yc^V~mQ3?84J>BG;- zT4$lStZ1cV=D+EahBed=#5P(P=0V4SBG`{$ug>F9`Ep9Bff|KU71P8nZ8GI>l-mBC zs#q|I@LJrXJYxK6vd(7Y@V^Ec$k51jIayAS5eh~tLPX;TbddmPvO|$UKOGN^0}-?r ziLD+)9D)~lHRsNV7BbbByv=o|#vDRsBza&3a$5(vBV5_<(ZP{b_LcxFyrSwC&YB4q zERUWLs+oU!=u zxmw!)w@6(6w8Aq&gS`XF{P^;4?WrTcK@}4!Bu}CA!Bmcx^&~^NW(>;8LT}R7$r-6T z&kyZksCKgnTQ%iQ5s&O(G=Li`6}0Q3>Qy%X_2*I~Qb?ROsIFEx?K^LDmmdD9WKsl4 zk<}GW?iy7fYyS16DbKoilER_cy?V-V-6swc+?l1>th zdFaYU-!{q$|EBCur)cm^upI1#jzAP6bGrow@wn$#u56Agr-kw_0p|X$iZeaSvY^MYiVlFgk~3oG8Eq`}DBR*L2Kwe1db7denXrinmGORTThj`#o}2 znve;+HZG-ay{~&%_=@uPaKAo}_hU_y>c_N`k@&G+O>NJzGv4_49IAsi3`hVV(~cdV zvj2lA`whvpQF>aC@8Rcb`-?)SV+a(hXNc~~xGd`r*khVCdR@qf@PD?P|_r4^|gVxH$6fbI|&sjpT9!=^Eeml>UO@ zi!Q^8I3lFf>v~?gD_MB``Sapo^pB;rlg>*i3g31qT$FTv#U0(3Vtf<6aUiLd+A0!w zbLai<4#sCZU?DtGg_>a3>@=)j+b+NY_w9h__-4%v1ipFRmmGlu6-%8PTfc`n45MhS8mbQ$Iotf&BVvM3hFq< z$p2nA%D&sy-F)XK4etX1|IG}kCjt-rJ+sRe%Km+_igf+vU;6fPYWI6+Xb=Bdpk@}w zyI`cs5m@QNCo%nRw;w1koP4x{4k~LOzkgo%?$s2#-5!nTVSC_Y!eHdf6teDT4vE`l z+O}=hcU*u@zg9+1=>#mjS=T!n@gLLWN?$I}M@R9MXL8x{I7DYsA#!~`?{U9U=dM$@ zFhO(oJkH@F750ZqO0nL-8JDd;^+rV43k2qBsghJf zYH>=*VpxXw50eoQTK?({t&)`yl)u?uzZsy{1^sxQ95Q|p4x>n0Tuwe5cz?P;zv%p; z;9g=JwR35rYVE3g9}r@C;2p;%#sLD%!FZ?RTn^;P6rbYU|0ZPOjed8Z5MYsDI5aXXgjIys#F&iNW#G}2r>@1^((iLs$faLYM= zp0;K9xsKV~r3%fYIj;PaWnfsA65a>u zGc!h$)ny=&BVZ~V7IxQ9#E3}ElFLSZ(=DQSFP@ew|IW z{4yL`dS4;6Az&)GG-nT+tX_f6r+*e&UE!$Hoc_yckVGk%0?=t_NlMvSUTj<~S^d6ABR+R()m9h^LMS zcLP!7A~_ui>Zvbu0wiMK(WLbvG>wfDj)Z6F!JGT`GK9v{3&t>Fq{liL91)m-cMui2 z(d$@G$y13TPQr8rKkR}YNaqHP4k!22JGS(0b|VJ##Le60jC0^vIt;FC>)xeJ6WtZb zVRj}atAz%mSqG2xhq8Fy?p6J=nZVu8r#QeAEO3vM^;!OBRbzO$p#OOs)~dZ;eW0_- z;A<7*@|qxcs|RcQNcb%}Ks52*^Y43gc3#}u9m8KY9aH``{d~s+zo$kiR6w8AJ2{k7 zWl4z1?WYn-J9Z{t-8KU%AHTf+fGSqY6y?Jg^}7N9D1kfRa6CR?w6DIT%L2y<@vgdg zt4)WK7lr70){veDdn7D>g0*MCfWd%k8P0<^2j6Cd-4%W@U7l>wUfyr>FUt9y0Lzt^ zHMIu|BmrS8_yc}x_|T_BPeT(;a68R;GVgcn_~H|LY0gl@wp71d(iX|AO7sm?;HxQP5|jeH zV!7OL*KM!&L#Fb+lk5kXOeKQYWv?;xra%f;Vvxj&Xr8NPUvG(ZbDH;Ez4#{K2lWqa zdcRlB=pwtsoy@ThJ{HF@@vB{>b?7v4sN>~bOgC|#e`qP5Hf)Cft*bWCbOu*(>q}j0 zYWKf$pog%EB>8uB4$c+lJQZ|rz^GCCgsAfKvZ}pXH zbfW*?r|P^B;bL{~!b|F^%Y2k8JKkhkEPTE*D%JgHY5I7SZeYQrNp9|LeC6)LpyzGb z7r1j%Em&S{e$6(Yt7D1{_)qGgi@V_aJyFq*M*TMB$~UfEiyP|2r)?_U+gh*z(I=1p zM)I5;y+3(l!YQ+RDiE1gq$QnP_@GnsgYCg0hME-5>pORpO&r@F;{#JV98qZE+1{2% z?~+`@WK{Xr_q~(j#& zhL)kC*={h-Gd2o@#wk@9MuQ+Ii?``-W`nU%xU}iO<7a!8pSy4Pc713z+3NT`gl;r- zHxw?ZA869Q$>6J$eb}aXC!>dXerE1qA?F#9Abx$yH5%$fs|SzO7ujU}zGbcuC?2*h zC79MA&}#0<FCSoawE!25{Lt0omGLW8#m}VauwMc9q(SG zA>}2@`+oq?-vp>R7!1VmP7w$XSdIq8^0ggHf)pJ>>P^$Nry>LRZtcE5H95FA(fh}z zgKd<=4ZFtivHgGY^Kr(Gqf%|Al)xW+c`TAU$@eJ;xp}jn1Fyoum?W#H1dz-D^#5X@ zicB6t;ul0Yj*<~Gm_rPIQOEW1Rpoc=mH;~(W#W_k(5QBJE<3EB!24otF!0Z0-fl#F2( z04$wq#|*3ku;EB{Dk+@nbG8SM>!^fJ9Cf7baBTY+sfhF4)w|Uciz45+;uxQrLKF2V z?&X0hN5PK*hiltGVp?km7$o)W!ND6XZZ^&Y^Cu^Sp8Zs&lf~I}P5SS2nWmZUuU=bT zHd2=z%j^6%U8MrITG(0*?Prgh4EXnmrEJf2OZ?$4p|8@c*mzV}*q;A}3|Ae6lAB{P zE^~>p?kM>ki@UUQwzZsT8)|lGdx=s~_`as6GB)*9W!m~V_A!}h`Y8BnG^bz|&24?& z@RdJADi$PD{phba&ZG~YMt1Ai&7eIRNqxQ0kS$b{i?h0zu|opjyv>^_R}8{A8>h*v zcCcc&5Kgo@%nxZfWodC@Jy5s5ba67-6U3R7hkQm?+5LZOMs&&a1AnFIre?41S(wC? zI#4Rzp0K0>wa72}TuCWKSP0SA8_MtZ2JHLys#3H5qt9ZNVY_Ij5a#&Y1g(0-S+!)) z_Sjrje%n@y%*{ug3U?289^Ifk=nnAh=Qb*So^s`||3Ym6YRu z{2$-wN4CCjb=wj<9E$XUy=SV?SiTT#c~*CO^!`?SJQ)mMwCX4lT@9z@n3gUWpKiK5 zNq3BdqsycN`s>oDh4h&cK_B%}?ZuFi`bqTgE@B2_(cPHn?%9YL8aDKQ_r`B{KMbJK zBEyVWpiA1CNIhM|Bc`)hq>}RI=Vp}UFz{4~ z*29HvKT;x=c=a|<-_2<~S-tpbx4orv=ov#k?4m#UWa>P3&)`eX_TN|EbFJ(BLDi)c zJ<{K8ALTjzOFFh+&|TWoOg??OEAt~nT`X^{_@zDDg?oaiwPq^=64hz@rglH6R4;$0 z#3OCecS_z$qvsk;H@Z4lDYVasWmTl-i>N?H|J~LH-(k<}SuMDEsftLwsWj9@=+9=? zLgZ<;bf}HG^3L&LVu8Tc3@4n%6YG6>R1%E9@@x ztL7vYjl^HMZg=_gU8)$%L<0Y%Hp+o@!R108ia+7^8+OO{|NXT_|4~LD^fW14u(3s3OyhFUIDsZI{c;_aLX8v zyK;tlGqxuus=Pgq`R_^rpW@0FO_FrZ)W{98o?#ddfnYl6dR6&5?80=FClS?6Q*d)d zkD3%t`{|nMIQsYsvA8(T2!xR?5|3p9KKy!(Y=a9+koBo(b+6@7T8V^U!jMxD=SfX3 zzXy#}deY_yUXPO#RQNdU??qZ}g?L8LYIl0tfzGn5W)?Mbk@cHLw~Ui4q-rBEg0DBN z$p-VPb?x>!i(AgWGClg>YHw!mkw9j=l~)6{D_Bu*PB@5VxO1>gW3xOz*=}u9`eXGk zhNHbQ_4<@`Rgv4%4G)z^f`+6y7Qat#W%8Eru0x|5${IeKYHs|*4ZnB0a=gX0VBQ39EGC)1;NdR%7RMRZ-;sn@isZW zF?ru}$u;cRqlS|Mc%z-g_xqbuzp~Tn#BFLWm%oELxj%blT6F_=^&*pYS@H}6lim2U z{N0I|SqFW3rQ6lPojZ2bJDm42&Yo5(Gn~kZ=G%#{XpAo{xT5{ymnd}Rtp9ViSE6h< z8LhKAb^>WxyYS-YU2#3nR8MwPri#f~^OVM-QCh!?tR32}#RnzJ4QL>Z1U7yk0jXD7 z#c?NkZM=-BvCrC{1zLw*d2aHb<)|`8!h%`Tmf{>hj)8ax@>wDXzYYzau5yCkaXZ|} zk#?zIkzz3_kp?WrM)XgcU#`=#>5*4HLspjJ#6d}>ns<3?r0b~7f@5|E`gOZek z|Iwl&@`;N{I-e&>PnJga10UK~suq3llS!;1iJts252-Pzy2t{hHkal~(8Q_uhB-9w zrN1M2Nm0pZ#p6pDsN%z)^FbQ$B#p8=OqE_&A|fID;hZ6>IRcLFo%W@C#nU9L42p!9 zKrj-VlZqYj!`UT14Y*14Ku&R%@MYk{8tSisG?#1}x0iqF;T~Wd>yl-f#uEI#?@!yz zH>2xcf9R3RQk(Po`rkL5>+T7Gjc=WVem4bWXZ(#H@9|MpH>OD@3?fW$H-I30&I3x^#o~2dfB-UcyoyYU?6;Ar3 z@>N=2*iCBM7-okXhg9QZZc>H)wpfzyXqW}(o+BW%dmvn-PNI#u{%ru~YkGws*RJ&) zGtNc(C3!zF=_ zoYZalt&;QkCt4p3prII%9FtMd$f4Q`l@~Q`>C%6upaakA&|$N_SqRR3tfA4*f6HSg zljf0BH`!zU?XTBWL3V#24Pd%hw*O?fg1Sw!m&Yv@u zuMK5&MU?p<=Uq#C^RiU;75Hxdru{o_j{?5p(L|YMXiEmGNA!sJFIb^{CC%z0fTd+- zv6tS`_ZPP3VP~E3y!9BTO>HPyQByMDwDqgy5_ z2^_coqZ~W`=nraQWiXebd^W`PS1o*$$(_08+;Bf@YUfEFa%wLY>;HFm7bcVE3dtR1?pV}ByugEl4eW|zn1f?#z1OXy`%c#iIS5_&O z+8noTXj-n7TQfP1UjI4h5+eDe700Q;ME&0*waAiU;(UN;4z$)sutc<^IPG74-p#NX zr}??Oh)aZQT)(@dgoKk|+BNIGre1HY`iE~4g*`h)t!B7H@84Q2kNHQhm!);_+4X~8 zZKBG60G!=}2fFmrq@&C-mqUT(-%H=X1FE?TJBQGH=*Fd%;hzUmqqI3kd$$Hz#j1S; zUm!U*keh0?_u0*Mt;E7_cAevzm)vC|vGq_{@Qbo%#D-&{jo1-$g;uz!OT2CW?eRy4 z96xZi*mqwHq`R*EB)D`~B^lGJKuenLmah5XGaxDxEoMO{v3{qg#_zhyumL)t%(S@; zyY3^cR}CHKjoq*2VwkkcF~9NKyRnSnWn$+|kJQsEB~NvDlVw;HL&|1TE>zcF{IP7e zpK+g>7Od-7Px7GK4E&dNiD|54ZD{S3jd-eq0DY%=)?sbYzZGD|DXgyOVa zGZY&xu9qkLZy$V9d&p-*SO9(NDhm*iRVdY{r&8XqIU>) zlvy~w;|ljUWdRm1gqNk;0fwzTlja~62n-k5ifcdmmP;&3Ge^$d>8MD zp%*K9pKt!0@_f4=Di*?Fc){&!*4(wWVSNeWtS3!S>p4eF%Vms(mQQLAa_V}a_ycaY zyt?4pWs9@yo1*iyxaf@3jl$duPhu#36jmTO&e(wI>}ZsvHWCzD(|mRvq|IST{RQKR zlb^`wO^vwC7kQJBcdXacvJWCqC3nlJE4+F^gI;%xvEkigCJ!kw+9Kz%_7Wu=wiF%U zrVRbIOe#!rj5JufFYolI@g>36eWUuu^4ZBEiuL=B#KGg!FvG2$l_#Oaj7v*$dtB{j zha&v9zS)qbr7TKwH*P=4%usss=)Va`*B74KdvPSD?fYDM|D{g%tCd?`bN><&TnX7Y zwSrV!U8vhR2ry!P+qASwsOmb&mblv@@%Q{W#Ena(^l{DibQ!mZJp?C@RYYuHh`;79 z7Oe&8fRW4#mD2bw>+oxyOw74Nd?}}*g4!QF;?gg@mtIM(3_I02+EZfx7`|L!e*D+v zS7u1}{mLw+4PE{uVq@q{fm1p>86D)~(^byR{FCh6ig9;HbR1iD6tzfun$b9lD*1Co zO|w8*ncIUnxjJ2FamJa}z2jJwvG~CI@V$b%796VNig;&!<4u8MSC8ylgZrOfoeY<1 zN;eF1%st2$9>qsBeM3fR3w05T81eJ5oPNAEtRMV}JU$6eKKZ0F*lx$Rr=Nyse6jPg zc;V}({{p8JwIAFw$EJ-mnC(}*$*+%U+4>KMZ1r`Ag6{EEn&+{Q;5=U+wC#1|Hw&Rk?Za>W@9%&6mwAo;=ZhkzKXrISRFTguk(sl8~TjX;1h%f9a1@ zs+(z%lWl|6O2$8lSx4&5weyL@`X_0N7YwYCv-M+KDiUwEhmucYej{t%Oge~6h-8Uw zDXDJ1W!WindF&+5*$nl;XCQAlO&2`X#0dP+ka+h%-STbw&Kr^OTcwjGb$7pAc@=Ha zX~sVF&)r#Btwec8y8e^qs~qk#vQnO1%T&pf3cU|O6;(U}*Lz$kq=?X1S?wA#Ny>(@ zHH)Sm+rA0;nfW3D8O4l#f5P#OQ!nqg^{w&zw=&3A_ixx-j7L5dg(n7!zX7{d%u6A_ zt8etPo#t)Cxie2=ogCclcjcN6BjUdN?I-mK%=G~1j~Ojp!0tWkWXdw`szIB$Kh z4;hXYen0y546b?d`o1#ew{)O(YX0S?^1GjFuECf8>l$_5n*VCCg88CIVix+(^Rtk^k7^ z$*etE!Zp4oPLvrovP$<7R1YJ}@Z5MFOM@lM4$Hv-b7rS8FUfFil20x<0BzvYTTo8y z3=)EiATskZSZDH+J;__`j*`w&-nVS?u0E)nzTYT&)9srghKQ?jBNnd7<;v6qs@97i zUXjzvJaTvMM{_aQr3j0H#Fu>}tY4=nOQw!LObTzA8?01#XnzZR_Vu7eXgLv&eo?$+ z;E4G)*?-tHeDjh1sF&2`E#^O6V9b?~xllo)L;eowMAt3%R#x*m{U5w>SO46ZeD!Y0 z>@!9E%qfTe+Qx#{ue4woka>{kE!C8*T999G7?5hVzVmPN%VPS~xTYRt9lB|FD7afX zrYfi%GylT2nDS+Tn&?~p*zDZ2ty9svAY_iQXU0ZkbWGa0Wm*_acpkIo5ckMTcsN8# zim3dYc#Q3dp>1ed8|Mqv%(nWe z`rH7sjKJ#>o9)SS9sHNHWmLZp@wfhGCg8Fd+L6tlBK>%A{(4|{%RW=rlRV93yZ*1p z`lz?Rjw0_mU7pX#pPTmJ-}31TS9^Km{^4QW+v{I{z8A2p>6cilU)Y>)mdbSU`Rbpb zEPvtFapz4J{!nlxE(ys*6t}8>gL_&3OPyCjc#oE)k~2QMnV*^h4g7w0VPuFu&)JYz zWbQM0dTngU-ft(IEII(~QzK>mGx{%xVeXBIb^CwT)`)LEZ?xV3cbj_$)YfTB%3yxt zQ*ec9Uo5v4E7S;EU$&CzF1@c zcVnTvK0ny_I(vsPD~*PW8>4i6H+NM}RL6yDLgoxk$-?j5Xz(>p>G#T|Yhc#<98jYn z8p$(t3884>naP~zJ37gfn9+i?9O5zfI45u#^D6p59QimhDOst$*71o{i|J!Ru6-Z% z_Tx!DYMWuWbWDa9YP4SN;anw;`5I>H9Ql)7t&bVE5C3P!&pC6U9fKW7qM|_MfW^Ku zMve10qwVL@F(+LoSiXRZzYc`pV%3S~%P;cq)+f#dIrr?X$QqL>g3~2(ZS-{z`5rgN zn7_Cyw&~*Q1z6~fMbyLvF6EbR$8TGX_`kAHR-R4CpShj`n~n~Dj;=dD_ffK5g84M& zkCJoj$t6)nUAKJI>?b<{LmmHk?XY}l{n+gq^s@h4af+rX&!I!&_L0e#j|ob{59>k- z3nwF8kG)LgS42llrn|iTR?ea)1M>5w?hAi1W%oJah!+s>B7o2x;!p2Bx2;Ux#@UzF zI8plo#m)Y6A|Bo~sz{Kr>}KV@%o8$4bAhF2WRIjq6mj;;OI#u9wCCeKai>rR3;h@M z#l9+7>~vOKwH)ShlKWtz|D+8*l79MJ-jxXWPPrPH6e#>iAY})YKHRtd*o4a#PBQC> zt9-@zI8MVK9-M4=l>NMCk2ntr>`WLL86~|lS=qgPaz^{b@`tythouaHzHTrbo7xiy z)i*XyX1@Gl{NBpm_r~{0Q$y`NzISD$H}4t<9~z&WH$H^d+bP-qkmKdo8!PhW?UWm} z)NlX#yhG00wgh{M;XA3Q_EQZ1DgV@|dP7yfNfgL{HCg46@!ByM)sJg)Y|7p-MoR7? zM82|+v{!+FA3M5}Zj;V2IL8dn{yDzgpccC?s3A2{`RKA{^y`T#QEiD6Zn3pDt96l% zi2579JKk>abVH>r>(oy;uF4wB;4j)#DUlyu;prEf-Y;i{YCix#(mdoAxF>_eTFm8|4sP5^AWnO>Pm-{;$BXZMJ4_`WPMWhBObJ_TY zrA|ml_i;``dK*$YPp%o5jLbO2a8k;AlYD#!9({9OZ`PN|Zo5?Wt!c_-W9zGyW=Rc* z@PL=avD^WD*RsD>_3Vr<-%B2kEJ7B}%w<($e;})(oL{|d8j)ChH`g_#-(bD_^r4R7 zUCoV^KX#MFA$%N5tR?=Nt|Ru}e}}c4IhuBwwDu8&lUPeXzHL|3oR{zoW|HJA2N0z= z6<{+CMQvAhW~k0fT!oaECI7jzEh0FtzIbO?ZKh37=@(=zOdgUa92B3%{95wD`%^y< zAsMq3Xj5SQAj;t@Dzj?iIfuHnbH49o@dQ+f@f)Mko*k3<@?hBgTzSrdlRw(-yum-j z>R9^4^$wQiLOd>W8u!kRIc61K7}9!Y5R?CTwm5#tuPhmze3LiS6!$Ucgrzs^;lunJ zsxtT=@?#<%7jSp0y}ucXD-9Z6@qikS*Aw>m&PpJ7?&!`d8KH7Nq8`ZU)-PmRBMUD_ z&bfjL$Cp~8hD^R)_+M|cA>|Q6?`@Lo5pDLmBwplguf)u$uokDryO$)ihP}{xsl6r& zJ69!stK!_(bnQJZ#GQS_&oXjBkny&;=E}oggHfLseka8g|9Z+B&KslF_C2I)T*{sR zxhwhc=}|`aj2+<)I9rYsZnYk4DHp|&K4n%B=9*z%HEkbKpC$J$4W1JBlWfp8D2YvG zT;qyVq8tXAO8r?glq5P>c0sOj*F01{9~C@&B`7rUjjG@{eV*22;G6iHnB=@Zqv9s% z_FePH^2M^ol}EwROLk~{F))1+oIzwwnIe3qi9kuusR>|99TTC``5@S@Qtn-teIBF3kODfPzADTeD3 z_l#0sL?$VoWh|*zUsvAWTVua>)=Lj%MjYrPLyU!N?UHRiw@IKj?Ae39P z_ZzbK+{IBMOGxRuVpyjFqVYZA{^0Ptz5!LYgU5}n4XZJZrx;F2K75?q9%Z!rX+qcT z$vMrN!oVaOENdb8ohUg4v{aIvmC(viaMyt(lswZPv9`vU`K=(Q~uR22-!SV;?-KjS}PXxMuoE_8A`upp` z3FYqH@mJAY)B5XuxA%|C!gkVrB_;B9+h%I(QBopne@dQjDSx+- zW&d=<&paD27G!!Wt}N7t-3!XTK3!#3jeA;ZB{VuQg78WJ76#ORfYn-=jIP<;m{ES0 zqY8x1u}prKgd6gj4p1Pu?#Kl1H~PGkk=^F{(y+B+dQU>H^AZ73r}w*d&NkIjjW>K^ z!M;*_d(f`NQT46Ws$ZYw&B+2_=C70dkGNZH z`2|rAF1e2vhOIQzNS0!9Z+Uz>>|tGwT|Yq=_0jAfReyZ?UtjsQy96uqc|_O?9exDs zmVet}ZmPq-;+QF3`xdK;uosnC+ABvo7tg*i_s(9J1Z_AS$~nMxtK4mGSwx@4{9ZCN zkmfye49*jG*M8{-PU~m_WKa0m|#P<-=)VJ+om2(wlwBZ zSC(C}qSWf|>t>PBU0#1SIj??UAN32@e-SgsTi92+xxc9Y`;4NlsXu|SU^D+jF!8Fn zS?C#7=Kh>#H=N%BT_{xU#~QUHIbVhD1Sq^dG@LKczD5#5L?8qecA1+T)!$2WlbWQH z1K-tWC&?xZvFi5N=vrN}I?emG)doy1&GNCjf57in#O-^JHROgB{hMGq zuTiy*o7)@+qC2NHA!*gqGc2J?$VieH!Y;VF1^uI?a;}X$J{N1i6#XQBxN6q@PRBiT zQ5FGUT=_QdWO9A*tEngMrGvt_=vdC616#F<{$|8)*>mUYj;|GSPH%in>8)${nO&*4 zRT=m0e|eWuK2A8T;qRBemW+q=Lq}^##P+2D=y36IwXB4=jUhF)&QY$FtIK6>^0DqE zSX`BObP6BPGU#?9-UvNqc-q)QK_{lFh|vhBe&b%#sK$pai+P zn->p#Vt!k%yyq1XVwowPYd`w#gG23I31`Z>#d6)Fp%((f1CgA(=1-krO+W&U;+aCO zp}SvE7$Xo4xP#io7QKU$hJsK3*0{E3j{RN)hbenH?yZdG1mq TeBsR_I!(_|*f2 zH}^)H^vo!61y<(bPfrr^{&d!P!mwz!O6Q;j=}MKI^TlffiVCRB4=ZExlQiac{>yCz z^5>1$24k{W%BmlU)$8JHN~0F1F|SB@=p@u`a2zMHY;_$=<4m}{DD zqx|e2*OOPGF_lVN9ig59)-^KwEdB(*gAjG$m4453k((Lto#A%C zEPa*i@j5*{?`1AibNbkPD=TZft&^%wI=|2(~N97Viw=z#5I=|ptu2} zl1l<_@7Erc3@xri}hh_H$d+)jC58&n0aF3$twwY|l$;`^jH=?yPz5LbGX! zlQQGa$mhxjid{lZcF}5vC|wK?!sru(5g$T87!^GQlC5v%F=SPHb`cy{Ekr;(O4lBo za^|n(JQ7o`Irsw4qtU@~sJH%LP_GCI4;XgS*I=VO_EN+Je^boPSqBpC1r-h`1GsBt zz}$>jZha`v+1cW0gGODb}7X5x0Q-gsr3`&HB^?hm}rVnMyCkcz< zQ=CS<)XioorqrsjqY^GjX0ny7CcEH4A{pSsAqH7Nm9*fagSvH_(L_}M(g6(z*==Ak zryS9)H=QF&DqIE7Sokrw#T|6q6dT2>RSYmFL7IE72w5Y{1Jz04t@i)APxOL+Wn)Mz zbPeQ&V&HI$(aI2Ohb`0bYefus2Sa#hM115GF}PVGYw?{aqP*KY_-B^GyW%$i5u0@~ zmSd%f|4nc21~|G+Zii5{w+P~A8Aj-XZpUA{T+<@{JqsFNOmQVm_cSi}9`k*=vNN$O ze|#-ZaL9CY(f`r?wz_ZYKDWP=@|(trQDq3=?O&M5q6EM>gefxU3>$-V)Et!H$&I6# z!0~7Wx9Kf+OpZI*3SSGsJJ40A?Ve4|?jNFkpF!mLh>87@Qm{&}BG6(0$aSVD(P)zg zf+S;lDWYJ^B7DFNg)1^S^q6)eClLGN0VMO=P8Bqy%oz;U1bGHSh630X%Yp~-86<~B zSU)XZI6)UU@J*=`H&_u5fm8V7lsL!xB*ZX?@n%a!Kf2i)0eXAo)E}u4zug{Ov%Cr@ zn}bx0pLbB&n-`ITok5Zzg2IY*;xt{N7YYjTD4QXUiQJTqA4`NGkYx;d5{OS-EY;Sx>X-hI99u z6V~FlIH+tSw>Y>76#;_M+^Nx|WCC8Eje#?6ixsH{FpbBvQO^L^hz@DUGzu}D-;M8p z9Y~xu&V=$MHr7$a7`iB^#R5H)bjA0%LeJc_&hjo~JF6nW0m4JrgL97a#d-D)Gnf%{ z-0YB&9DO_ZRdxsr&??Sf4JiY;k--oqR42X!GQ>cuz*Aw~h`2rXB0&fn*5UL|V3{c1 z1hmO0BIuSOS`F`A^n3x>G$jGtX$*;Mgl`~#Ef6Dx$$_8x#ZBGj>NWxVGS3KwXZR~| z!5*Bg;!^Z8*SGu#!=!=4*icW|y)0byC>Ffqv-Rm$;!>6JCY*`myvnc+Js1bTA_qYO z?x*6Ja~T{Phef&+|I|E;Bn*?`BA6mc%&NSiR!+Yu%#Gml(-*XhC)tY zrf)&Wk_!0{nG|mhXd)w`nx8q58?=FKgt1cn#Xwfd+#DV}@D`qtP_`6XK%%YWum_9f zI)|CEQ2dW|4k)#=$k<^un_xZQ#K9cFp~ z*MEKiJV2@YUQDFn@c=LZZ;f$-;Av)pATHK84U9Jfo&e?~auU#}S`osu7BP`Zqlm#c zdK*V>>RVjZb@4?~09Zm_k-CLIU&cT*GlfkL<8X1@b^`UX;f)2B2JNfm1!6C?K1P9T zfp85BKg5G;e+cu2Asdu|gfb}-SYQCN2yr}=Zk7SS;A6xj#e41Wo6uD(*QN4RFdu*m zfCpkSz+6z;BSpM~GXK6j#u4DbfKodLjN4PDi>DL>B~g89_`g>-=Yc;xCfTct<@#YoD!Y2t_5sI$TV3KJKlOB?1Tm8uO-C7HkiR4gD6m^i~=JQ#>wt;J}Z zMGndE@dnW!07wTOaZciDJq~#lPW^3CDCk%aD9=x_V@?N!PsFYe0RR>uH+A{OX@&6uhsd`)$2!- zFfYX9L0+D8&u*;T#XmhKFG~}h{;H4fKS|O&JlQbQ@e<05>+L6jZ!Rx3}SUlosNSM z_TlVcf<-NW7_5Wg0OBz)K3L&e20$1o1lFJ%6aoNpBoGCPRRIARXyR@Fu1zt9%lJh> zZ%+JWyf=td903-=U&Ty8xPW^`Fs27WcmZX3{%o#BMT`?*9fb42II$`-W?G;i%51Bn z%5GQ1#`52(hZu>m(4JwaEOZs-5XOLMj_x(oIj|Y~lh*+9srwn00RsVZF{0QoCLpo1 z?msZ*usKfDI4jVIfIKha)F)6m*R18zca0^*Z6X1O~jlmY$ z+)~EdD-HA9(LC4lsn5MZ1S1x*5f@aFi?Ldex_JxQ8|O;_W9$&@j4&qd@dzH65eO#B z0W0OB=zw$qa47xq1lv!fuEvcGxRv>E457b-(x;Z5?ba35pTzF(ZaaAWar^ygW##Xy zx~lMyspk7g^PZFcdd4Dl@FhQ~+X}b7ca#Y2Djj<{!{pp_7M_IOeR*x=c;nA*=i1Yz z=)DX-a0U-(6$2fDfoN#d*;z7<3q%XUlQ@RhX*TX;NHrTuPmXnysmvD*-4w+C#TP;e zI896-v|olxMcw$KGAAa7?&ul{XjJEc^t0pK>CTebK%5CC8WhjfOj5h(lr4!#z+465 zpj_3cMrh3eOq=Go(=jpErd$9$<@1NcuxEaiNNWtjU0D$Y2QZH19J} zVZ0`gUzZy0WpGkMc>m;wRTVMFm*;QEBH8X)6Q}vo0XI>sR2IMpfyRK8PO8fQ%2EaL zmXbBBsA7yw3A9cefFWM53_z=<>sHO;sdt6>D;xQS0mx0(BqoM(zQjt!qzBf-e#c~8 z(^4b86nx#ivJn!^T}RW1EshMvKrxCLua%^gWPn)RX+f(5tO#5fKh2JJsP#l^ciIld_Iy=n-*@N=*X+8Ei^6npnu){{kDPp(uk>0LY6Sze4D8E^$xDltG9%Fy?czu`q;iSSh`{BJml0L6p?Jpg+& zrOV8rz&$nzus&$j!qkJAb_qfhbpjLiXv5mDZ4p#2+Rgy0g$FPtLa+dLUqLT|9^G9a z9=b}w4D!x#`~<400_Ip&Suhs?ig*l0i7B2n;}S;s^wJQ)s4{_BmCh-G&pu$t5Cj7_ z3AY)j8ej(~`#6kdK?6W7puJ>Tm?f+Rh6XVk0a4;OkO&UI2x2Cfup*g&(s6KRWLutY zDhZDqPD>_{VMsk-=Ku0Odzb4QJ@1j_CN{`~l-}erL@r&Ct|DfU2y1=Wl{$$uS{&HB zNcTv^!T=~ZyTS!dj1=Gp41Z4Pk_4C0_Xm z5ohI^AM(hS;h;Rjp=FR6d2Y=j3kxrg1PZAtyVhyMv+L|hMhS6o5BQl=8PZh9cJ%-r z;1yyPa9JFLWI*ASMD9TojvRkn_ITDZb#gMsCmRB0i374P#Jo>hF@B#f7qIcjD{uQm zx$b$;hFOoTX4l`?lRtEo!kdT9ZO=51UY=dKHCDK+b2k4%$uH;oQ_kPMzwf!Ff(mwv zUI715nXS#ca1+Kk3;@)^#wo0fuuRrol(IGx_#qU+RbGY$T+_ww%;l;;fr{#$w-J~! zBa92Hfj)rL1c&*z5!`(<*g@Au@D6YUN8f6wCdPybBmSCQ%}yc;o8A=JN5G&MM8Otf zVr=}Z6eKV&nIx(iFVdUVMITUM3FdbMy+?S zZ9N|Cf`>wzORmJ(^bG@k0OR{r;dA`PE$je8)&K|_Q?(I*OvRs`cZQwBdFi^p*TXm@ ziiv{Lm|%ohpqRa1b@ca<#B{a-rh#QJyZW7_k$l(E>lN;PhQnqy@-Av zV**^u#Nm`?3GG& zrhMp+J;@35wA?ar`84r$movCft@r8_gN0?Ux!mt6GAs^JW7 zlQjtyfiHpX!Tn%E9*iOlr0cXN(D~7W@)Iy-Lh;1bLI_fga9%gU5K5U3Mc=y;mt+jG zR$^0V#IvAtjdb-Z6j5>CMHtdRtT{aVN(C+Y_t$Tqw;S)LXoU4of(c_Kp7EPnW*+_3 z8ZvfJw1BR49<$w3+PY}7v&e*ZZI}O5Zt*%~wq`)x?1z@Buo-w~w-jr9v%5*F?(Dgc zo8lJDx~{h?9)9+|hK?=bxgI2QI-5FeW`0P3yywVdsXsUOed&x{H#Eq+;J(2|`{QZ- z*7TqCLX^pnchwWl7w<@K|4?adm;ZO9{%9z4jQ`c=bqzMCMw124Juu*z9s_IV@75p3 zp(p>H_WX85yRR>6Zf$1y8P@~OvAs#KQj;!g5;JsRSVf}+>Gs$c{1c@y%oy=SIrL(K zs+a|)h1+gQ-UZ?Ulu6zh;*{h_>ZReU&)VV2hO7xOG5w2J9r+(EIZ7D zW($=5Gp(4C@rvqpj+ed>9Q5dw6w<_4cc@3-KHj{jgO4oPe>KQ|ymMn`W5z|!z7qtJ z8wl_sznXpR$tiVH@d2jr<;~5T%G`114Bb*9)YJQ*@}Rt=$Gi_{=YE#H@HVdb?-ZXd zEciGYGB-HEb5Z5Zn6qQ`(0;;oK5a+7iT`GLd7g6$eG|00b<0YMn|ZK?6}>GK9A>i> zeVZcizf{Cr)kwa_!C+Uk%=U*1=*XcRwrPHXHb_ zBP|Mudj=_(gn8S#A$cNVoQLoJfAVX##^JM&lkdy4wrgm}s+KI1y$bWgDUDeZ#R8GaCL@tq8wIOuL(j*XU8 ziHtol(zWGZ`0{wmBPhiF=}f+P)8t+0!rEv4UG{LEzdEaLsvWgot`-y+QS9@$xc3xC z0fA)Sx+{{*-^IULykN+KYg02y{Hg}8m>SKzAF|4eEtF;-qaL5O{ z^yCCrIZFSYd}(ucJd(x@da&TYgLptj$~D=%G%Bb;)l z>(X)Ur0P4cpTikvLW|z!V)4;#=_ffkhC)xCmB4Hb&!#!P{_v*_IyPy#uFmJOvGBu9 zSiBfJ^!wAJgc}{z%i|>S+e^x?uBoDt%M*+V*Cyr?HY>6ZN#BNVCG}2)Xm;M}`S|rB zqS0e2x%@r*2bl~)!q^?zxcciI8Nouyk`zg=8Z*{fk@muxkahu)?q^;OxwXu)4SHW| z508DFU210B6Ww)^-`6=h664(jZXix${)iqMC4Hm63t))@O(;MqEinadWQ+N<-ANt} zRky#g3x72)(v}s(f^nAM_o{-m|fs^REI}wN7LHg!VX0 z4!Xm{bDKhfsGhhEOM^}Q46x~jc{25NwTlMx)@qB-BI8TaN}2S%4fRrvQXi!@{mQ?o zE`eMO=MgW|kdY=jRraeqC|)@UPl?U%I8<8(mo6yk$!lE>$URav&pKWR#v=LhPE1l- zF#l|T6bJv=^e~coi%F-NM;5^iMOhY+WqqKrV)n%SNyXdi;iE=AKa2@ekjS7wCp}(h=8fQP9@vtB^M--$-Lc6bMp3g@stH&M+ci!Couei9frqN4-Y-mE}Dq;~o z8iGHav+TX%{b9K3>G>MDwoxe3N8uBdS&oNY42eBdQwN#=Bae)`Of-MASt|^~P+lO9dU@gBo=iQ#{#~}O z*YjJiX!c0cH@q-Pvr)A$9ck;AF0E$YVXg>+PQzrDHr@PBt#>_t^Zar6 z-tgGk`AJyC6w^3qp$a4Cd}+9k@VeiJ0O*LViYfhP;@awX;q5smu*kQOD|i>+w4I~B zDClxBd%G|XrPcgJ_09aI*7{Wgw~!S7wo{zfW#hjQUX2>7im)eWEC1ns>iROYKq9g6 zoq0su&R}?3qDe8!)X%;+6jPeZW2!9T{iiGK&m$jDmg0e;=YBx**PfIVWOKVrga|A= z_x(dqv1N3V4{1*sIk{VH6`ifFLbQ@an1W zR^63(8Hl}q-0hyCSvV6dOrnIbAV}X(yTO}BDu@Jqd@# z1GJ#hZR+6QzFyP)#362ww5yQDD#ME*W*`y|y5_bj$^cwYL~NVL4J3e7zN(Hy3;cG^ zx?B&v$w}}LX3@T%s7B=WZD%kXVfnB$iEu=s!a?!!?*glJM}ktS#*}D+3=X#3>l(AN z^s^`3r_1^$d!W&#b>`;+GeBjOmkpmm~#iahN9ddzP;z zmfnll!k^H8A2yT|0i?eElj_2g>FLWTzAe&%;f$2FrGZ1A!iU0P^uy}OQc_Ix6Oo(XFKQBj-h`T)?m_7w~YJvi~!P%%%@yziwPb3z!MC>FLx(vEeu`bkTa$P(W&H|z@y@4@e1L+g2 z6xHG3@tx=iR>_R%&Ry6Y((j>$Q%>elxcPwr;FnmY1WdFKO6z2YvCv4gVO9!=2BmRh zV6Fo@15hCfJE5UWfL6xO0tgP~rE-FLNn5-~XXqL+@kun;`c==G5VqgaMuBji6#x z(}w;JO3Ts}f`dSB4jX>OvM9n#L1|)zFp`$x<%UDT^|%$?Dp@|CPx*O#PXt|8Ip=K% z1SV55g1J!~1>~4O6lb?ZqB(bd7z08Pg@>v5`H7k_@yjs*1c7iUMhV}j0mZ~&YmtNN z>%4KvX)T$NUt&+{a7IAW01Yd*lcTyM9Or}1nV#d|i;CHhG4-IHI|u(=Tdr+0car(H z^!f4gk-Vl?O4Inqp&0Ck(cpu@?2fLcP#pkBvH|i#0-u#0 za*t7`0>dcZi$ZaWD{8S}pq!GZ33a~lJDBl70fU4MZV>=u9Kg5IM=#H3?*Q-k(`(Vn zW0+lqg(?h4p}+r4G!o>g)+Bo?G#GUo_y|9!7A6>RfsI&5)l3Z4NFGT}8OZ`M$xbOL zEr!1P~QMyG>{ts*kVZ8{O} zz=DRhPmqnkjM$LHii)4EAk#+ z;Po$;kSQiEw<5CqciL@%mxhj?abDskzfHONbjj|8eCQ>0vG5+D$ zj|$oVp>SK>vQ`ZN1?%LGjwv5Z6G^=!KxR64rU&dtKt(1IoQQnl7pO2ESVp{VQ``sZ zJQQbZNMq(>HySSe08f{YX;0$vN3t;eqLn07wL|=z9#6$Sz(T(#~mJs$^RNiC-4CASE zCvL!bd~=xLT|C0;S_C~_Dq?XYmL3>X1;~lTmT@)? z(7=vUz+nU-J7`_(jUsZ`*FcydZXiR_g^L>I4L@^9jFlCp0#h>K){|xswn%f<;h)rVRkda`t~Te-{3JRYbKrp2n34BLn&WyN~qfd#Q0P)EL`rqzsV4= z1(<0uAHbwac_SzYR>E8f<$|av(elJ(?3MMF;RxNP z1R_Xp;W{RmOvA?ur&EMM@^M8n1u-vmtrMiL&L6a;x?|BQ5Q?SFiG!Wy)XDK(J)EG} zNzeb1{okAO4R>FEib)xi(>CVnneaGTp13>}>0J8cmj_o--&eOU(F<1X90|OxvA5jAY@m?4x~q_;9mn?%&i#Yjv2`q zskY-}H1*~Zr5?zGaFQa z4D*#mY&fi;-bOIch$~s~o12_&vXWQxFg05?sQJ^(lB)tG>SKgMSJkxL!LJjWJe4D( zTV1g~P7sdl-)i*KH_!#&%7Oa#TGWfancs~q{Mn-EUlk$GUc15LIA(1IxRl*1=S3!@ z_|~pd<&fzjARDajo-U6uRwS4&&L9yWOtsT+pVo~BJ~W_0+%vv7QaeRS^tu6yCl2&W zvcLkBA)Mog9A)O#Ibyb3F`I&{>hN@eAn;7#vm)QhxuH+<5vKmO3zasezZy#TkZqmVl9OEUlqK2o)oc z3llp;B>z0KJ@XA`N73{bpCw;7+PHUw&0Mi8c%p7-^faeipVuvana@1xM@UP{)qETq z&z8Ye#EKbutv2n^L}*U#SGQaj*;r5BlW9tgK_=ftrA=9z=oland4qQjDp zZLVKN;g*1l=)Si_uQ;0&EZ`F1;(GpxjASr_oLtW8T~s&x!HiMb zpObcH4^T1;O^?BBg_6%=tV8i4jDQR<($O>@`N9B9>RlCE>qn5e&hh#~x{MH_dXIjx z3iCKPj%KJ8s6H26h)c;?oZx2Te%T~%^YL;SBMM``2oDDMewgKPLCr&6@+dvdlb%I5 z-D&Pen1+B889-whig)G8(9FWmPW!qw^HZGPn^VMOwcA{t2-I)uTJ0Q%S1q5kA%-&g z$~pVWXPMG8)&~s?YGh@EfWvY9{>HrcA0Mfuvf)0>$+<&p|6HH<92;?&)ng6{ifumU zzO=xfXSoqFz;?Yxo8<$aSMz`67v#dJIceOZu4cFXncMJQ9`(CiyK2)WZwh*8E;Dy! z+FCjWW%lS?g=;@t)+1=f*i<`J^jy}hyzn$EeC%Uc6VrF~nDs9M1@k}LTow7N#WT@G zQG&@bxYvBCde(cwIugQ91kD~JdI>pN}%+j#!Dt)rpszVU+P>2^im{x*~5)}(vcBIwkP|NF0hTGT!VxmdYKzM)#nbX zaJ?Wy`k7L7xC6z{&1A$Zi*NYTK%1YGgkuzokjRfBvVhIFD!XN?#*f(--W6F}-j13# zv&!63+eqL!@A|Ut5D`w+)Fbsh393Mj->Pw?dCqaGF)ciAALRW0fpKWqW0Jn}=MARs z10xHTUef}7=U-u34_HuHtY$baE?Uz$X~m=eWDnkIwccD2HhdFAOwKKg?uIW;I#(BW zQ9n3k5DDj~HVvqD$CJ;+0R_QB4y3kT0YDBq6r4w#47}nwxHru zR$HKuuC>U}BD>>`PbME<9WY2-AG}jJ?0qJTh#sQOGS}Ww4*nE#_T)=!>_JTapPN^v zj_b^BI`VU!ap_s^f0UuF_KlOx!E5U~?B}Q$ON7UNU%m?Q=rY7SyehOO;2_rj)jWnlNYzos7I@$pCXVXpXZu`SR~gMp7i81YVe6OQ=GlCncqJ{b z^yN$b3k^i$clCTndyzc86zTypY;*(M=;b99GziOrUU)1wqZumKf#V-SU-p@^@ymRR zvY@J7LR9fwVyL%`8>p0tTDIj1Y=H+=$q%%A)i{ei%*_Bk?eKqC`!a4u>s(EDoruWf zx|m)8otw(AJT&**_19!AcorpSo-|FR%q~7fqL|q>{>LYvvmwbwcbG7hZ!{;=DCt@` zm(*M8?9?0xY2HM^)HqV6j~4CFfT0-05@#@!_w0NsAg80V{QFX)GZ->2>_}o|FaP6N zv6AT7Sf6MwQ?JQi=u)s$bN(Tttr84CyKg-9S#Ebls9YwjBT?*V z8UIhiiyj?Ku!>CbgTl`k@~Rclj-MrJLc6Z*8C1zc2`{+%`CB#)l)Se@*d)Bm;nEJc zN~q*#o`a!)48MUqb{eE$T^N2uUARRXMK(X zZ!vtGPJG0>PPqBH*v{@XgLGWZNGg~2lXwij$(_}C+nCn$l6=OPSjoc{!uCq4k3GZq3J0FVlB*0ZfaZd;Uaf zg5zFrGw(Z1d#m4v6p4Crr^E*~x;efW*quY4;b5#0Yq=p$Ajnj$d@i@yM`W=Lf?P8W z^e!ca{w^3|AZiCxp@Z&^j^!nd8_iujAg%J1mtRxpzITe@q~wc;}a0T&VdCAaO8L0ZbHpN!UomQJcPc^napWaP#bwhh@)^xv) zjqAU%a;FQP_e!(`nDQ--ea;JP=dAg4K4I!&%ZJj^qp7tX4hd{2vYZA;$@{n3P<)OO zF)cmVi7tOP<}GNF+{F^l~I7!f{B5J*cW?^9hlZf&MmIf`Cc$`@~qO zKdLXU`d>|u2t6hr+<_F&A1{jrP_llGji}G}Qst$-$dj7d$+N9L z|Lx~%9_je}Nc02Ly}D6>41i5nXlD!|B}k8MSl%k3h86BbhWrIA#AsZ9V4gY z^S0QfR@{I`Nk&{I^2MIqE05siemWU$=%2SSbei`s60?)ZSMR`F1#vedW!RgtV;1_H z+OD!~h{eLGrOK}C`C+p~=C93-Z!ZlVgDy&yDh)q*opj^mhfJhG^ITUz)1nw9(I8PFU*c0e^=bYVD#;Dyx8Hh01l)i=t=1Yw#l9brR7e&63 z)?Ht2dvrA8T?ay!)SU&y@bx7c#B=MMXRbM!^Bnr*R)=amDSC3T(y&0?<7dyw6-i~e zq|_69;h1f;*3r^x*6xReGI~QbS9mh{z*qgXr$z_qChmcbsZ9i**#{SE3!`}E1uDx( zUZL>9hvOFVCACtmQ4f*B)WY3QLFDgyV!h99Yj1m;56P5unmNUAQ;IbrO(%*;=-V2T zN7%_>j8AToV6~k8*cHj?O5Lf{Oi8W564R3kq#*IL5Inmm(=J>k)=He>nd{e3RjzJ+ z-E#Qmr`jcie0OF(Jl!MWljxtP_|%#c$J=*!W~9og_ROwt=PiEi#9 zbd_;YKVNrYm!)CpC0%;yB_)+^q#Nl*8WfONx{;QYM!G`@=?>`z=@bM7{3*gd%m3L| z-_O39xpQac&b{ZJ6FBlTA-btQ2yd~}@voI*#-c=HWbLbDxqVs9@?x(^MD|>sZf)ME z_C((AtIQVDr>#)Rd4a0i-`|J>9p5m6IZWK9K5)c9=oC$f#Y+swc5^CkOm*B<6~+!t zGT$t8^(;+OonskR+x$16eW6zKa}+>CtFzg;&*BWVn&T-UF(p5Q_#FqVT@B`0c56@i zz3A2(u$519FP5Eo_5z4Om&jdmW#sR$oM%_~A({IsRnc3hWrt4wv)ILvjnLSmoiEsb z)jjrQdpc~*eNaAp#i~5Wv|Q8vCOD0cb`PqV)IbPYWip}h|$8d zdup$qrjqPI^GDAeuj9+K&a3csex5;qt^33IzlkMDW!L&aJ=eiE7$m(jv`>odRVW>W znm#IL#u0f8=HTVBq@JXN?Z@_~%SN#DYA58;LPrXe(E8>^gR?Boym1Tx>SJnU?#la) zdrq-Mb=L_Sl0rWMmdD<7*W0-^wm*cl4NX#BhkIL(XD8hF89S+Qrj~$rE2nKoCHeT< zwgfsCy}7TF{wN9We2F}#>(S|Z{`_9>zqU)LUsW%r@n@cIH^rV~tIDbv48hJUT=a>f zDwHh36}^{#th9jdh&(*N{T)D0YXwI_V$6d)Mm1t=<&-Gw5dlXyNQovQqCY&#%JJF8 z)$RZ%nDBqT%BpIRB+{?`5*>y#p3D35m%-yzuK1@j7#1n!)Rq(;=L3INWqBxZ=2{nYfB16GAxM&cWC;qbXyd_1=doECI z-0P}oJk!?9KIJ}2XN#5SNJB<_h1+#14MYZa&x^gMc05ue?oSTJNYL1P2_>uPfRPv} zvgE-ICAp^xAc@OsYtHXYqWR?>Kjmt3=nB0z4Z(O(CUSf(T9XsPA1cYmg(IvU!pq17 z;5U-SLmxs8xX&ZG@X*dv(Dy(zQh+wB@E@Vmr)++wR399c}oEL`+^J{4V#ll>th{(j;l zxOMgO3(~XnOgHo4%$wL#r;}OqhdhC?a8C?y8PdH8(wQ z-dU4+qfHhC(?zCV8i^ei4^CxLmLOV{2ro=VHy6qlk?h|;UasRoJ(A5-%$a+RqU1Fh zE;s-hd4sT4CMY=-!emXi!(q{jJbDJSs<&FxX;^K#im3xRy3?yvx!h5-LvcBi!#eM# zO7ejSh2Z3*ejW}iTno2&rnF%O$C8(Uh1ee~)HjRNwa4BR^9V@VEZTr5Z>s+Tp`OsJWd zN_>hCgom*FCrvDb^L?;{7r%6LuYLgNu@jx&pDb}b+4khk$xhe4(c*h&P2a2OY0r~& z@^>}sC7Kd6A0stqw3g~QN~;lP6DMq^NC0>F!MAG@h23+HLQ3y0zm&$}qqJJDZ8t8F zYBxjX>O9pl90F}+Lb)9=jMG9<5;E&iE5STZRvw#an%#kX%GpI7aI$KWXTqYMbVpIO zo2PA=woX-($Q+l7N__lVTA}5qfvrh8oz)Q^iss)}#^&a}bofbASKs-k)a-K3W_|EeKeuBK=PpdE3&Y;#KWJajFIvWSq=*9ZCRGUPI%+*tH z#gYr+>YW;To9p--E{SuPP8E)+#7uL2l7>we2uCujDG;|c$)rj}mI|Y})k94+9RyWW z&;cA_^&P{<(|V;KdWwJD76v{08`?M5&grQSIoI!C2Y6MVfp##*ctL%fs=dtf2PV<& zU%ZSw3UyQp|NX2S%tyRH%g7O6t0yzRj_Yu98daCgO@4-fflznE6Yz0X=6^gBR^V!n zJPb+{6f~uA1;(!4SR{!UrVNQ~=5Os*4IKt2D#uBnRY^{x!>H!Be@d-Fs3&5E{m8{Ax6 z3y|WpYbRsx&KRqg01;kS`Ks7OA&E`Z@SC@AofH(-`3us62|hB5&spnoyB4=yd}m_M zM)x!LAv$|QA-%32%FiVGFnK0e*YA3t9|$P{-ee92uv|uUz7S-FRj|9|s|*Q&h&N^y zIY;*-fJ7o2s z{dag!Z2)c`$!Cy8QVioLRJA9ZhcfJ!hJjz<_C!krd`R*i(C7Bb%A7zdsJ1t zP#}p1-|{cy0y@(Fd*A!i>(`6qiP$M5#-`WHtBf99ylig@tqD4I+=3zI6j7|tojwio zpN>a|dW>PMky$Wt>_}<}pw7h-b7?@EItH&i`FdRdA5dntt2Cdiat@$Rznzpp<(it&yrrXvnNxT^MI;htrptjZc3He;U- z;=1M<1tmAXU*aJUxmxP@+wJ_RPaio$ za(5$N_bYDHJv(SNl0H)Y^sl^F-On)hAbIe|Q66b`UdguA9cAw5;f2Y|?CTBR^e5-j zdIUQ!(uJsFDdO1<#NNm)mv^nn{I|(N%zmg;dF*_s9sEPJKc@PllMckqltVl@WAt0u znq|iMelLd8@aFn$i4^pUHEpk@0plU=en`>BZb-(uAd5UZiQaefx@9T{r?wN8w~9@# zZ^Lfdxc5jY)5N4(y2Wt0i-}SIk7M44%iTWq@ZK9UV$Y49Y+rEacfoG#w&KzWTjB7_U#fV=aT&HX&#JdCs@ERJdfxPYZ{Xq$HhtJnq9x}QkA=+BJew~HMcx_oC{(VdI4X4q=nGOc75v9fLp#$ zO_o=-YE0>GwLj<(6X~(8T1}~k@WqeRoq>rI~CD&KZli{ z?2J`}S`d#~@w*muobVX0R=dUAqW#%rJ)dccX+4NjwRYiMt8uza+d6*&JL|dK4%Y>% zxP@x)7kcf|?A!2~{US~tC`<@p3@mvyxlG;AB&4J8P^jZ6Qd zKmxG-(6)zc#qMd+b6dFxLGf4RQ*3NZWL@4QX!x3yQPOU6fw=#h6irT6?pW}Y~#0+ z*XlPu;-@`7MGH-k_^j}Yty(1p#L%cT6S;PyP~`1x3w*-dd3utVE0}uY*Q< z$p@}BdDcQnWFqZ|N7dR|o-N%wfK4TYQxYhOu$CT{FILOAv;l&}%Uvq|$1icfdG>&O z=!@2WhSwRtgVCZ`I7mxp5^4ULnfNxaRj<9s%^h-MUXuGx|0g?>fK}3!V4SRnNAwM; zYthw3pHtCJsLu4^1WiTpwlQIvxsbo!Wl<&51pkuhf-;tPJ~0dMUu2dNMo3`6(?RaG zBw?LU%FLz4pqI4qQ4wQu#4{(Lc7`fj2ygbQ zC@fsoOHoc~T7I;+Wks_}{^XLcP!;#Yb{aK70t{282tYT$L{=l8kVh=u!0*$W)TwY8 z1T**(F3V+>*gER_+vvNL#b)gDgk2f?YfxV5n_|f3Vv;dg+!zQ%0F;t8`HGdFnKN7! znd?QuA1c}Y&x%S;|_P`y%O^$AD-^rhUBu zT|o43=+LjnIP+A|r7h^iA-;Y**Kn4$-i2QNZYgZm78AeV4=pJ_j=ne!0LBFC2pMs7 zFO|>df&QX&S<|0!&?*D)Q?cV{`ker2_8k#Nyz(~#q_9zD9L#6SA2=Or!m>>dHuO=s& zH3_HP{vjaW>7_- zNKv&iDIpL4#L<(FVLC6tMpc)Am=+7$IuW{H2n>SGXT}WfsYhTw)*^k^z{K*4yqK9- za8I(AGTZ*NCnOC^EacbiO}QnpYTWY%6FPpS2kZ|w8aftlEZ+K;DhDrdp@Q%}u3JU9 zqbnV%Wp$x=ML6#anCM`1`gTZ+yHDE!RNo6l3SiQqk%LU5HaKA~v9J+1EFfC`UPP9M zwO4Nf8JkQ40B0y%5SUFk$0s^GxvT0;f+=qqjvY=(Pl5@wEK>PW?g5EZVq^`KRHVoT zTsC>yL^aD=3{N zp{Zb1>NplkJR8C+7#I+XWMWak9opqVf+{lrs>xfQo>tH!97`?qOfy}{v1EkWN{6WQ z&Z~_4AK!z6>i@ugvnJn)Udm%K?@(ua46v7`WAfdbp)I8o`#z z;$hIkGcI(z2$`My?~tHOOX=OEC2~%lUF|f7DVbSrisl_F29wPbRz@kU)(oWp*1w5G zMRyCr^YN>5klf1y+lw4KWXLB;JN*KOw@zP=iwt4$&f*mnXCEDx8_Z{42JQTZ3`>6c z27Xla;Gj3ivv2l#DJeqpk}u@JG|*zoViR2&MeDn0mEo?Nw8uEVO)5Ir?Uxa|7shye z@0^L7q3t|BtFMWrsg{)rm>~jNVpXjOh;b$5sVt1YS0yb@mJJk$AdN$Mfc4%riD8IU z&ci%b<9&l(j&uJ07A*kqD%eo<1-mH?7Q6ZY`@ z7wp!9PbcO|sl#SGAimIlTBT8#BHoLNUjq-MNcL14&jclJI|CD5U9r5YCv9A&KkR%f zwNTZHiRQycyq8H3%HjjHu~UBp=@ zmKaIJ2E#wK>Sc$PWVuE$8PZsJ19d~}7rJ?IVE-D09N8YV5k@111(F=D?1~6QLJ)}d zNv^UYW_@m4TxwK(o(Bv^EA2|v4k+8bpkKF|9%o;vAbya{Lpkf-mp-#!UPG+J*A0;0 zX|r9-D*{z9Wl$mEOcmf59w&d$P~gw`17EYsrb-H0emM*d{@8QTnu;^6>6dc9X7GsR z7~7xVD#$igf@2z`-~Vk#{V@BnHYOz@nD|7Fe)a;*6X-YiLrM~Dp9$jqYA&D758{C5 zT$*pIUgL4|!2$B)H#jKVh)HP$8XQpW9xg;MHSLM8EHl>=N_KJpejE^_e&DmBnApIP z!(`Nple}XOVmuV#ZE(q8ocJ`OuRH=y_=HA8QDcUcl0<~kU=~$TL3?t8E_z@lj_WX{ohyp_!J$;p!E(gOLL{>X!fYYEFv!8M%?x?#oJthT;I} zKoRgNYsR5%lZk92Dz4!gg1VsXU{c|32R+XD8(LpUtO{E7?Bk~np5sWKO4tIvjEeL; zv^2plw;#eyQRzb5%d3s zJ+%b}W4!Px{IDRT^QeCXXrL3Z*yIQ~q!pckFNY=n=9(6(LSl6r*#)##S)W0r!Qr?- zLK!VBrq(roJvtb05{H1`6)C}UXy^n<+%{YqCQkxY=ceJnpw}va4g=p|!(fpb$}#|) zD4Pm%tA-brj@(kpTK-OxD&CrAlWjP`UpAH<8@Q(yqQg8^m!Oa4~=&Ozak?bKj`&e@XZk@ z4`KEY@W90v6r?;iIuIDE#D9bJ?6_(8_BOXj&o(8)pm10TR2rX65!+}Sy)Y7S;KW^L zdcJn@)gdH9g9r0^kmsmp^FPgy%Xgv}VZ9`)5`Bz}1iYtSavsCRCp!@@exam>x19e3 z*QHT0M$rt#KK*?YA{UKTuS~+_S$W($3z8C5Xk8&qvE)Br>b#FFw>>vqU04DO?$>$` zNCm|TqWrA%YOOFdKgM~^zL1|nplJlOM6dxm=G<(YNBcgy|heQ#PfB{tq5@1=; zr!BdFOV{rKGGf$%I#EvqTJ{ao)nq}y{mFT_q*nHHa8C)P80UZGsX>KOQBdgWXA)+; zvdP^eL?z2i<%mL3b^`4rVUJ{8jq! z$%1G95o#$556Qc>64X;SZh7(RLJhA-&rPkB`9m}f(kp@|K$zBePVw97R>*}HX6yT- zZy11ZtmOV9hHv+ebDxtGJVfhBP0&02M^uo7vUDG{;K%;weieoK2^bB%(-U0s_|;|i zrWT3a?O^?q!K@YpV-T{?8}I$mTk*H7urCGKcL>1`4`V|!h`p4VP@(KVnuuJ>|{<8q@{8{0S=D6Znn6C^~wUN7LxRQ^4lvB?U5H`LJ3$AaK~*R z05tdkQH++hC~Ux7trmD+#Nx6s;iN`HresQ)@G z1y?~|mrCADK7B<2bhr22YLleq#7I=kO9R_if;$xeEwfsD9L@$ZG1lI~IlIL)ir1bk zoS2+8nAz4akc|rAsnU&D3Yx`a5UJ)PO^(x~)&cif)4;LBWJ#_V91NAss0vW6t&`2* z&Fs4GLC_KgzdX9-{`b5_cw3A8pD$iAzLh|6b!^-y+?BH7c1>tdW90i`&KKjOvSonh9n3}xU17eL^LB z9m2sHrU(wlI|Zu@Qz;u4jWWPCY3&(nR3IlQ1G+(S!EqXfV$Zq)h0ktZ!By9UhRL{B z-n7-7lkNXT%heLM@!!0?DdzcT`rI)pR5hGnEtNW!oIi{0!87I`N_Nr}PbrD8Nj^K| zY!q@`8<5?{_f5&d85ysEA~25VQ2ANgYxB~!(#x*)JL0zj9SgYURc(0;z=s6ah;{^1 zMSZZoC5y**YKX%|8(Js40`&P`E`@9TFwy{>OEGWqV@MRIANjfaJyv)LZ--YJ5_3rx z>OMnMX2TOE%l)4>t4BQNFQ80lnftW?6>Uz`N$qEM+Q3bVW8Qtx8bY*1V>F}ZjKTG7 zox(Uu(jzcfg)Apg^@yWj44u{Hus716q*9>~ARxeT1u@Uar-JP-eJ9eY%N>y3^IKXE z>Ia^bVnbmpaYus6Ur~QJ$MiPMCJu?6=W4W*IS4M)GB#&OrUr{t|F>SEXn$@Si`;j{ z`TTt5a;j~N3nNr$Rik+ep3sjwt*>>_3xQh7 z3Ho!sH0LH!k|3p54c&TeSvhoQq(uXx$aSF>BI)49DmdV?vFXgnfn;V%Yk@Bz+a3h;o52IrR3KH%UW zkU$~=AYm?(STh4VS~fQnIcX*|Jg5wcB}vdCKeAJI`jkoVO?Qfb{w?Kq0(oXBp<7^F zk)%lY!^JaRV3PnW!0ublJX^BA5wI$PnWOc!hi+DSc{0F9%Y_+{_xIJav#*DHay6%w zLbkGu2^jpL;XrJ9pfSfR)FU>^BLz88+QPa8UKkF8<(MaZl!zzVu&R+|jLoa-f0J1)Uh%XTq<8T*B{H*PK}keCfcP8K?GN zsdY+oU-&IPYjFEYa80?ua*NJcp*gs$2`B@>Ta7sGZ}&(mz4!n+3JiG0ub#N^b0vNq z|6MWIWMN`V@It?bHgUq~+kcx<%APBkcF+Vh{)_?7PT!qR!37-h*8Khq_(_SbF~BZ0!(G7_+jl)IOh1+h1SNERIQ`PJLOJx=+uU!Ku|I8x zzO3P)B_ouim`6utpiX^{H=H9!l^3uZj8Vi+s+j55(p!?I>dI;&e;r`zz@bThqTCC! zdVk#>kaD>4^{2a<`(eY%3L#k37tQTR&l#fpHM%%tRmy zNHs*=>U{g#nhGNjtD`wKa4yofN`k*?{aLLw-YMu@8&LIqqEAT#&w~RlCxACdRah74 zC>MY~V~#~f?xU@W@fEGR?Ks%SXUhBK#l)-=?*h3rP(u(S0$I3+8LAFZ6mZls0$^Yl z&iOu5BkwP^VJ>_JSQakm-QLv&1AtggRmMZZSeTq1Dg>NZ2*Ca*#-d*pr|3yXUvG-l z4qUyzR}c$B2TN7nvdt$-p%DvHolQt!b!LujfR3n_vBBdYS8aj?`jo%VY-(tphhm~F zv=L8btq!Ld2jb^f4b0o_G^i|>9U)h#f_cSN)WBw?4~9T7-DUq7>;w6ZGnq*TGyK#{DeVg2GX@NwOH z`O?sF!aTlf=bhW8a@@bsj^VL2hrB314(Fo_juS@G*OICg zDPN95CVvh^z9kS-DpU8o=8Mq1Ty-R*<=(6;E@=NzS*&WRIT>}V2= zbdAoR8CfcJog$V@XmyN!B#epd^wW+~^|ZenPQj{cC!J5z&7nnJm|a(K`*Oi*>k+;> zF*lmiG%-tiwECBzU`UVYRaVsO>yRMvfW*v0*(UvOZ!7&LkFGM1RB!K~?tWpXdHhzX z8q@_VG}L0EZ=P|wDK4Q5y6`2BUh=vP1ws@;lkc7b$P zFh+#(^(+<=aCBEzYRui&m?a8&oY#Ty9@-DbBBz)M-R)<5Dn&*+jI`|8*T3gz6SPJl!T7edoNxWVoS0L~EB_#o!#(h>7nFFh*G=i>LjnKVWBf4X=p;@`n&4eR84bi`01n$Jt%B2$qpSxG{< z?=xkB7qoh4{x=Owur11^g49Ad`0ZOMBjl+7NJOvylDl-#ShNgAB$epzC`&jeCMc6+ zux*7DD;ny^i4ElXj3#vwsaxL8$brM`knp(uYHwa$PoyaQ4#LAj%Q6|Cd%VVB=9FAg zlE*n{po~%o!zuSO42$I%^CwT*1z-mw_QOZHAVLiWNz|0`h)Kw zRj8xY$EHm-&#rSbB5t$pk@Txrn>;wTDd|o1f6wtf29}ByVSq~g=d6@ZUEE$ixVN#@ zy8|dGBGSn@Tcd(h5u{8VvsY|P| z4Lh{!O znpA@X#~TmP!|{t>io2eckzzTpD}Rh4spl{ruk6$B7et*w_2Jbzumk3Udo;GT}zws0-(;Z&!5Nly zl;L-^0t!i<<;NKH`hdXEG^_z*%R=3!hzj+7JbTe0eQT7q3E#k>{P(qJ9)k~CGw5@kG%y0iWrK1l$zlrH?p zZ;#4fw`=;STCM8r*>Rz5nT5;M5QQVc8g332D{Mr6LzKQ0$FpXi|^LkYYM+HYALltR0U*Eb- zj?ZaO&tR%TIyBX+KyVS3Ldx$0yKW1-W@+RlW6G~vgsU00RJlrB9ZKSF^ne`x&3!RG z9X8KCoD1x!JBwQfgnR(@P5x?vzY*l!YhX$vy zK?G)&hh^H2cgndih{zAy#1HVxOi4ga4hIk6jnLjUCKy&8mL~wn!%5#^d9nqA-Yc?+ zvM?}`$;#u!xe!@}G4|Ftk%jGI3}MxiL=`*7g8+Qm5Ds6cIBAyZEDp9-gYphCHOmMu zm<&&dT^7+W7BjqJ9Z)i;$E#y+d8qZuku@q1V$Dc!|%0P@7wXvJO0cz ztlR`IIf|1UQynqWOv?CD%QpW-7Ws2*6|45{pKo_-ZgP7GJ7uw0p$&k`4Om|!fh}SJ z92JfooBj94dpUk5^uEX%+&AQ=v}Dg+0j$LiUY+p=(fslVI!K{SPn%B;+Y+q`z>&87 z%t$Kx*ZetTeVy?6cP4O1>D=A7JJbd)h)2OX%<`!J<@5vUhU)E_T!u3r*d!0NgfiK@ z2!2w()wrnstX<$vqB(#1&F?p_n?CdWQJ84EOA0U;NqLtwwyflUes(30av z4gpFm5m1}m#z=GwLxHKQ1A&fOimH~E9jk!;AmBKDiiDV13t?Jh{3IYKEq~TzFDxPv z0Zm5tl3Oa=GGRhLA+4lU5)|3OwX?b5OgMlj@B~#Bj*ex2Q7&ba4g~;n7!#PK%$XlW zN`lk-Xy;w4=GRU~x!Zt)CWQ3HAb@E`po9XQGRKvNAyeEGk;gfrlxRh24BS~UTuVT& zsNzcmOsfwj0Kp@rDB{!(s#9Hdq0Duz^Lmn*i~!%UG1J?t-kMG5F7Z-fXKk>BrS>1= z02CV3apfk6MP?nOqQ*bO`JBetp#6O$b|E=`6dSGTi24)nLJ!k5>RiS?h26yCJc7z_l z^2#dNJ?edN-oh~jA&9W>Be!FrP3B3p?l2cAKDK;Vm<(ni6vU2zs8?iQF)D>>i#fsQ ziI5QEXq9qaEKM(Y_uAT&bje^0J zS8N10P*G(|Z5;+OtZ%l@Nx}v(kTB?Q7^7v${_K+YsGL$)@Ae$bwsA|>h2T$SbpXKJ zkn#&@j^g4;X;jq?>hA<9B(f@5>Qp;2BD&}2rwR1}PZ*~b0XC`Rf0Cl0WFQ`9;T}fe zt}_z%J7u9K_A7ElN?fR;mz8DXOvH|s7(b*J7FuX~IYTgj?hO|ux8on>&Wf;v;Av%v zs@ooTt}~PyaAfu>;&4XnTh5S#g1~xDpm~zW{jbenf&NiQf;yBK4xoaHk`~s8EdUVGNay!+syU|v6@Ehg7(l3GU#dVMFq}=rWgZTN^s(aA!3hfX z$ijdLmY=v^ z5Ry~klXHtU{?c_5WS#pRCWkp-Sp@PIRMMlYpPhUsh<)$FO}Cd#-%TPjpEvxns#@8ubV4Wd`zEp2-4khWqV?e)(L<^Jo7lxF}JPF9gL; z`{JEGTk*vc3?f{JXo*~>gWwAW>e!!(@l(V!MldPoW(hzOdr+oRZJzX(pTO` z7}!V+U`l62l*JnJT0Wn$X&3AF=v-w`GLog%i~Goa1&yIovF@h3>ImooP1+p`lpPQO`%&`)bT56zX?Y)1_j+E|b@ z*6R8j&ON+DC5B+Ohhn2?+`pkj9-jTChk88x4juw9;ys3Ej2%?3^bC zOE?W{kCeQWa+j1`_-kgRNp)psM3v*#<|Z}|aRN*c$qcr{95~o6@}}$%ZnlSBiMj*3 zCg`uhUGaqi7yeqoAkS5f-RE`s~Bu&ooHs3>fE z0@7Av7fkuzXf27X!71GY$^gT>8JNM}30MHN{_P%n)Sk<{S@&}j>uf2g4f1|wz_ZKI zbEIj+oNZUurnS|}nfUMLpImgg5sRO>v-3hLZp1#g2jcY6j;;UvZ;^%fp{d7)uk~N; z*MqUGTMG@&@=OPvj3(t%Z5OZ

%fOaPbyTQv%HY2-7b)qPG}C#WoggynqK z>2-u%GC5`y!a(w)lf&=lB`cV03)$3#^aVtsydB}@_JIjRP!MBY7!;bQCtJkv+H%hd z)49q%vR?xXz_e0OG-0#<*ei`z{=^4D9b>~|p)DM63?w#>jFnce-IwW&ra~hp7U)Wd70ec9sr zgu0fePQOrpS3Autuh^>?kb*_=As6=L&Bb>XsK3LD|6HbLjK6Zid7sqRJpJ15znNW} zyJ%u`;kTF?^kcy6$ezRd(EC3fPcZ5;ejSSHbvgm>IdwuxYQ#H$qKV%Ohv!w3CyG1LSp@d&vIrziAeqUrq<(lK{KaWq88Lt^X9YnD5S>^H0 z{lqI=VituMMHKa_{$~6&ry@{pYsz-o$&{Ntbpz%GFf(ZCU^6!Jkq_}z;dQ5G?*H6~ z&Xx4S@Q2Yp{i2T{rz2XDc`Hdg|L;Kd+q%)qZ@+T0QfE8=E@C+a9<2Me_SAl89LaLI zdpyTFp4;LfvDbSsHeY_dM4^75-qau6n+I7eewWC9hx+mSF&shNt!GWG{6M)bj})NZ zKK^los=qk-{QK#@L$O~dl%H{R9qNZ+kX#`|_C%vT)_#E(=rBgCn!OmJiF z2dJt+<&~{-X0M{-Y4XS3%Z)oPCxE0uF&SR3MV;0CEFhQl?on=FKw-cC{NTL6d zCEYT9mR4cxLQDGjw3b}lqjXLJM?ENUJ7KXktK^H!%>Mo#o;ul00IankTHsE)e;iiX zBUip>pN5@$O)xD6zc<9g3w2@B8SDy9H+S9SoBhaf`gW&_5kb_CZ1NC4uzSnQ?tC$dFP}Dy0;FoVN@lp<}^1Y+ZO~6xZ z2TC(PFXHr(H~zUmG$rfN9- zW0B4!`L5^R9m$NcT3hM^na>?^E(tvHThm>1z1G|-Oo?f@T=3^7%;3_d(tU-GYW|uC zAv|4K;}6-0Kq5PRE~eN6_t{=T(w6ur71oH^orjNI_A@6q)?WiGpSCClRZXs4- zC6y=+r7d-1c4%!-=6Lol5f4yoKs}YYBsFUmF2^3HOv^7SQbu~0@^mv@%vq=X`#|1B z{wU#4ylKyh{Zr9%4Yl(NIm+mwm9)sQU_-G__MCrbFVaMFZS6gV{6W#3jn3!Xg0AqB z&`YU34qub^DPf)?OGnXRS}U>`NJlTxt3#C+`10B|*q-X0Wu7Of5Q>C|#yzqp1RGf< z`cf#=rjuX+O4!rD;@dw%LxDxf&AB|ZCZONoq|gQtnh!*4gN5nehWuWJFy0T7ceq}l z6Ln=s_()ZjygiWgh5U{}GXV&#vp~X64#vY)akxiF%h;)%YLyL5B{7i#3T@fAB>Fl9 zuRXJ4Pvl(PRi_3k%qo44zkWB;lbbh`i<3&~K3b`&8DRf1TiK`JJ zB9W{gxgQwNZz#-0ZdJF5df;?4-y%41`t$y4!4unm@fFXTXd3DIGmsc66ZgvjANGC? zmzT#7y)7{Isaq7h%&}!6R0q1fZO)QDe@*;VJlNnbHBUnIjz!4hn3p)nw$MCy;8nQ_ zjhFceIMnmJ^R~pSypnTSp&#U-x1n$BNsTRkuj1UA8Hj{bXGX0wQAaAhd+j8vTf>8^ zl6-RHo-jo+ny~oB77{5M&oIAfs{_5jkt6xdDp=O$p8xf^nUvQo>h4oqxVDg>&IlS% zg-IdN84eUP)TaYsQh==VwE(7I(g|fb1~@z=0fRH%UYG)*|8!5d&-a5vq~?0?{P}}^ zqX<2?tA<~G@^Q>7Y^F_#27Qrw2$L%x?(O38l&CN;QX!pZchJ>wdZoWJ*ZibCMgy*zL%@oxDGOS+!qJp%!iR~6Icz&fjcyK?s5j9 zCq~%c9YyD(C+TDq)cRo7Q4nrtU1@s}kQEH~!}6?0AK7GA*0?;(KBDp@tY4Aa976i8 z_T^cp- z@44&T4;#y}yO!B@{tzAoaL>HA@eXEs%22^MGUrhJdA@4)f=5mvsPBBd6>X1mKa*^}jR0>&(!2C-OcC1n$pS+(wKGTh0S3pt zj{p6|Tht?n`x?vs_EzhPsa_bS*!uf zZ6GJ#LRhHf3x~-_xPZYOQb&9fNrD_Q<|9$OYVfmId5TYgb6KHp{S zeEAsvY{({1qTg|bbQa%jIL&|G9ie*u-nCP8CNeOR21(95eEadap!&zxtdTi-at~DA z&)Ao=`kExpoJausugm^^-Wxp)Z)%?lX#xlcU=giIKuPWxiy4Lu;N|p`X^LiSt_+8| z(*JiYdafq$RcgmP)s*H9{YcVlZH4ErOM{s%rGJsXQ)5r7YiD>3j2|tE{tF<$w5Tc> z&MajW-?*3&NE+n(>`Y`PUp$4Cs_xlW*I544$Ix`FG>Pk6l(thOvm=EnXf8}}*!n4D z`e#1ca|QkTZmxoMXMQ!{+WV~-)k;>K_>i;_GRGOYc~B5b7u3-+Q5K)y`L4M&gk~cG zh58+AFC+hI*uCfSh=iPDwU&&jmW#`pr<4WWiq4NY4(^NQQHClm;nM-wugJS>W`{o2 z?s$5XxE z#`pMkks}Umkq%@T4;DyQf&DWv^ngR-S^Ny5fP zDYmhQ!xUgo9zGO?D2finip-K>3-YgYbhJw*Mkw|V6ZnPW5cEuK<~Fy!nAA%kSYECc ztfX?^_{jF_p0prdky&RWm=#mSBfw7U?f4>hr|tCU7< z!n64U{wi=#rCL3{-0K->VgZ2BAgiSy$D2a>H6?A>)f-tPGfv&tnU8>IWRa1{A_I^g zEN+RANen;06#hSsuEL?|wu^3Kj4=j`9!TeaA)`ao(cO)pw19wgi88v8cBFI((k;?G z>3mI6N6~6K9FSz%&`#jIBb576vvs~nx89?JL!tCTWepEQ*(FAi4TKX! zRvG9V(MRA=&PR^IOj993C14h}@bE-jW+It-BNP`tnUw@8#Y+N#6UFQ?wHGhGpjl_Q zjFi z=&eqVk$_c#7}ywV_Pv$8akBM_q>9K%7&o+}_~($ABPFtW!RHc={}^T^@a|OR5I)I~ zBWS{YXVDX+(hNzXpZ0!U@vo? zjstzL`EXz#O(2+5E~h2D9O!LAE^<0ij_!Cn2@fCP33oraEgHsrV^Hnyt1*5z&g!4P zDrfpIc15+0UK$=HI4P;eEJlY|ua*6Ha^Fv`ttQU0&sr8GTSeM07;3XVWup(i57!ep6+JMg zTBFa}c>cA1N!1i#re-&*?naZ9urwfh@Zb89$k+P+T6v~8gu|{D_heQ-H&Q8O4n+|4 z5U+O&B_UCeLsgU2{#r}eQz=L%bqxtSgHjbu{($3;mKS5#WOC~a2n86akRST%ce`mP zR(M?0_@h2|oJBSj8Sh`u`d_~1Ksm^L^V>+HKY>pAa;<&XB6Pz_2xX8bz6F)#n@bl|0E$BrjN743^cX8FfiK{5~%$uk;PSdgk%}q9cVw1~XN|uJ43>dEb~Q$i}c= zn2#!drijx)sZi>`x-nHO{=aYj8dxY98yIT19p6NJxmC!E@D`9UB4Qp&5`(42MX2-# zrl-+YEHohHF<4Xi67s0AA)wcn@$mlb`ygbQ0e}Rcn6PlU7z^OcafoRtnbxsyuD~D1 zI?W}1EG=|$*`6@^5o_LaTaSRlirYSNTgoSQpK|zTH1^|1bx>fJ|8UpYP}SO<_~D+TC)m2rP(P?qT%=G%Npz(+|EDuj#Q625jj6+2<@tOSb=Gg48^ty zB+fP`-=h1~Sjmv!|DKFN*8v#$6n;1)2A~A6Gm!N`b1GS?X!(6E^Tz}w&6db&YmM>D zhO#}hg2;*lh_mDL7gnUrx-13#JGz@-wD7tUtlocAdotvDs!zf0jp8DDML@o(c6 zNGf`B>(!$`RRBii0gADxcCn&gEbLnlqrOo6r<0ama-(cC zq`Lr-!SIvVH2&iA2bH>4jDaB~PROIr-M-Ka1i}StYTIjzhA43R!2~|WD@w1QRcX#4kF^xu`}9$ zLhqe&19_84&E&9saq3pk;f*m`#dwB;2fIMmdz72MJh@0z_y&GIt7h11RsQ+t@n1pu z$#%oNU0YV@7ZweZ*q2w;F>UtA$zl{Q)QFQ*!?fPsGw)ZGC$)Ei?I&+aS^w-}i7mqC zTu|2mA1jMq$NZtZFL9T6W7^cse39j^iZUJjeMu&J=*SvU17q(<^Dc* zU&MoHZNG>msX(nu?6JNiWN-@<1U?P`61NF%xS#>*Bdhn!7%-NVw$Z)pdglCybRzDE zJ?h2bxt+VYn9Yj|WIn!RBS(Y-O_x^i>NRarctaFt7R7LL=3dF(tpK6Uu3xMbN>8kO zZSOd7xfEeV6ZCZ%^6q9xs4iS(FI(Imjs`c1#Oz-_mUHnomm?5CdG7<&PTPE@KAgsu zzI(Ojc8lo|!$YIW`RkaUmb$SN3fZ7GMfX4NvPdvU==~VG`R|@y&a-#OGX5FK;dIBj z0($=c@QgGsg%m7?pDQWGFecOvA7Ye}FeEqqCQ@AqU9y+8w^Y1K>6hOINZm}=Rc^SI zB?1ssp|u6jJOnwXpBY;baYTyWG?i)ff1Y{VXH%&();`L~dBXeW`PU=WcYg6oI9!K7F0q*CP-d_4hD(@`FOdJBV?ow>aE`J*m1&ULXR$>tegX6wx+l=(^z)L{b zZC~CP>gyV9SE$=Fj&U~8u#tyby~)|A;6GICKnZOwO7br6U07m?relsWHE(XHRle^$ zy|2Z@vczS+T3bN1x|2DNMKD=#)Rl&h-Rvz~kpJFW5A z{irA`W%FT4EN?4z{-d(Z-(E4(=L6od0$1%cH^XbPj{*>TgSvkkn@t#!8*0XCtXXqPpmm*R zgek2o6v&EBh(t||fFR+rlr)g8=ULkRG}p)Oa1cfCBmi; zC^UC(+dgs5Bg^m2E1Nw}G;fsC4977)aRq5y!9%g29<-3)Eq3ByuF7D)@?>Zbd}y!;Zllsy#{y}>Vi&x8+%k@Lo15Rg zM7{buR$VNAL-QP#`#@R0T&xrK^d`rjvA~a3H~?j4dH0@%_i5(s<+-i!u`fSb8tP6; ztp+t>3JlpOEsNcAS=k|oEN-Ef)@R$UHfw~&fK28sDjP!*ECqB@;mz24IpR4X+PuNP zai7Thq}Ic#Goayn+;?M5FyXFBchw4y@0EQ5Myht=p@*N%hhOqmPR){A?=>k)c6Wa~ zaB6?NxpYV7kHqVvwfwU%-DsvC{`z$^p+6QKsaq6yYLLY?uGqs8Xj39J15cX!@@gK4ql+L_6f!sE~f= z-1k&A7OKQdG$upO7}WBlFnGt7xcg1yPe5_MjV~O8H{Kc>SI|R9FsQZ1M z$6;>D+Tqe}IbMIm86k<2II{sVQ=a~&n4yv7R@UM{QZ&87e?wl*qKK9*fvVPVH?D^s zaxxPp9;Z1SKMc+7*chMC_=@}(&dGnw_>WoVBulyGdohK40B_z65Mz2|YKPy9=8MLm z6|v%jh=$Ku?`R}NmxP{qcYi$j_z#|ZGvxB&ADs|3OaIY1iPD;MD4R#;*T# za($l};_c_1dh{;(_XsL2nq1a1Yk!ofJnE~n^)MPH+`)Hk^OQiw!2dZ_?3+~mQn-21 zEikCK7OvJ-v8Cwb#aA6D&Ua}%bfm3wyWcN7^_4sWQ(zC$vweAZy)3W?6J zzFA7d<`8>&DDL&Nx52RrmI`QlOH@m)mRfm+s!UUMZ5jVhuGSK{oujH@fY`q{zdGKH9UaiiF7 z{wR}p2Q`BdvZzNlWcYnPFPVl;V4`?WiD@QhV1NWL4XTg?2RQ4;NZ0& zMjv)GHsNnonKP>9&F(OU1$A_<2=<=X=xDUETUpDyD;)YFDn_ z#(N!wQp&mUAdwqUQ5NO)lV=a*AMgs*I^H+wJ<-XpW~cOUX*_ToFW@{UJqZ``kQ4pa z`hD1bvB5WcA;EDiK*w|=2wybooGW!dfAFuJ#+&um-P;2C1y({sx5Y-|EBD1_w*Gqu zX2;`Gxl9a~Q58pfDll|D`z~>>TVv=5iehL z9kV;OBf+%PO_1v3@xiRrh5ZKthw#Tq(EZl!`Ya!G0S6*6zho(2(lk)}5x3s^9a$8z6)9;d* znBF-j&vWum6P368Ci(8c$f4N+PiZQQoF67hTfrZU!#rkV))0#jwqn^CbNfbNVVNhg z#_r#IqwKeV*iF`NTUY;4ppkVgL0 zilpvRd(E7uSwT@az_CVswj%S!oroEYa1waS+ngjMmdee*Oeilm&P=uH{SV(quJ39R zh&w{W^ToIRGr2H0R!X&i&!Kka1^XzaxO#=o`dFk3Y+E*a5<)KrRgo5pikv9TiOogW;)hjzc8Ouu;d?&E3ior;n5 z=k7;RHUT~Fj`RbpKQG6P{tUavvt=!GYl4Grx@SgM>n{yo%Hz!3mLKr%T%k9&rlR#m zta`rS=u4a?+rMet6Ms`_6m?>eztCMP{3&}5@O66gax2S*?k$i3`nfZ$}jXH)Fu{Vprkn#fZZEG{zAvL9dX%&(I;0C8EX(7JBKxsSz znv0rEVx?w?R$sIKV;ZW>RH2Q^WCoLUw)R5m+v)y%{wi=Qjpwheg^_x?{fd zt@$q2_A3=NSPwfi>yNz4zK2MML&*7%qOtb0|7#li!6y<&*@eW= zN6i@MbWyY3OnmwpMkbj-j?+GD%r%voTsZe=wI>#GG%NP1Qq;X2kA<3F|2=(9()DBC z*l6pa$SF%=O{<7?;Kk7|t49W`35!x}k}~!WG$t!|gN=HsooXG$pdV|<&)?7fXYo(U zZMwTYztUg4_i@4B?ktT-qd!gY7j`uy=WGq1=t+dAn8@HcC#`&Prj&RCKhtgKyQaOCJ0fM9sU zMidWZU7h@{6Ys9_$t8xOcEcHW~Se!EHxk2YkMLyhnNAJLg{okiTherL#+-Q zhc09vzW`lT6?0;>M+&h>90JalD5j{$!+k-OtCsqd>AO_D8Lt$ zHJmVJ;xPLKW5qxg4*)2)Bf%grP+O?O{=B+69E4rRSwt~WMIm(5=)_X1=x>PGTEs`# zHHsy}cuOOqii3ufc%U)L&_S+d04xc?02L@rohg%*y)Qw6`|rsgrcs5f?~87aET7g# zpI)6T6B|=MvMn~I!Y(ZzvpYQhMYZ*3`zq|u@=;)W7`2b>qrBxOZ<;Poe-T6;J^B3U z^BvzmZynzFn_pdJ-`M{}WW2h&AR-unvBh=2#Ybaj6w?e_cHCKMUEN%44P2vv#O_T* zxQj_rHPkjk?hWiKSLY4lKB^8oDV~=lx|CCb21cN#BaxPfjg1lkpdu9cT~B)F39WTB0v3~kqzyCRtOKzK8YQ<>wUR9o zoUpN%`E|c;OykWp=IA*aZDOG45K)C|hCS68K&c=CEK`TYOhos^Ew3d5-r#=&+-B}= z49z4n&jbCdkOoA_r})hE2Uk}++E?2|$b`c84gJer1+>l8d|2bDpWmB-q}UKzRwa;+ zW=!7<#G)YKh&jUx|Hi`P)6cw@;@boOM(d% z6O#n;)4GVR6I-#VB!ee`M4fV5Ajwe;qmgnYDjY4Z&yO~>PW^>axy>O+t~9Jkq_gEm zQGx(J|1$~oa;x71c-aC-`N5nhTQ!guFv5!g&k}wfZhISwJ05M&f(ruPASn1LY$3LH z$hkSiVZdVql%Qy!GQJ=IrvOU!7cnN0;k*XEkmyi=OoloFbwUgkwzKO9eG$lhM@qP> z;&Ckvpvcpf5#mhEX<_Hbi{n|wKmnKzAW_#60E_|SLM4$gTn?(QA7lr3z!f00TsV;{O_0Pe@LPdh|}o$Ml$6FcMM#59TiT6u=1( zgU}Dcjw3M5WXj+fb}HB@38$k-anx^VWDm1l<%s~xr(DI@Dkw7O8LseDW`u+;x)haq zn)qIVQ)^IJi^~BzB(5;r;Esb(&{bS@@gLL)v>8^&ki}qP6vJ2Is7b&vDAow2_-1~j z;0ikxXdT8~!$YP?!Um62OBv))wU{jY#pncN!KhFG1uC)Xp{}{QMwhP2Chj2uX2VCUGNR^ z7<2pWTB_2XTMQL)lSqf|B3%qQ2k%}sx&VsU%3CfRc7L4}1H9a!$}D|qy%o- z5yfK-m&ct5kYkc66I@6L?#z)?OR6JJ38=J;5e*)9IFDIh#YYuqh67(>6{Dc8c>dV5 z*ho>lz4BUopESNACOFPE311x{hxJRs0V8IDb4+6`V`RmURDiVzSphsnL_9(i4+R-U z6wxqq(~_qn?eHW~#SsW9q%etV_!tn6j9?BgVnhEWfg4Im7Tlmz4YqDr+32n~{uOdvr0Twt;g z!6JP@KMs};8iYO5CNhtC?rom;b zKSMBv%~63OC25HdgJbn`ZcZ{{)smjJQNaxi$CLd0Og!0p2miEh9Bp`(&m!G1@nK0u$Kx@tGB5`M{!T52F_oAMCIE0b2(%x} z79*Sb0p=pEf$R{BWHV8;T}w)WAklbEAsC1r48)b*h$kQ{B-UgXb|s8S)l+L{fe;u8 zGq1E=YfhaeBq<^f>j?%UvAoc95B4@5m#X2rVoFR7fA3@Mu?Xc@$I%t3weS^!@Ud$d zf776!?U)c%{`##Q#^<|*wt89W0w)a^m09}^gJBh-SSQVwx+>!0E%6=PGm@kt@)qVL zWkX^(F}4Y~xClK!3=I{T*aKX8xv_50G@F`q9%9Y3~JTonAV}DVXMm zd<<#}BV#U$9NYi95yG?5{;fXbZFqmIRgY`s(cZ$VDm-R5rsrPMdEQN67_cM74H_K@ z^fG`#?G(R9MT8Ib+e%~?mjA(^q6UiSr~%R8wteBzgWDXuxgO7)mkh^Kw@G2h5By3b zq+oT$ZAwPGZJhm9d?k5>G&q$stq(+@s0d&kh-ijT66i@p26XUXDqBtq0|O8bK2^z1 zOpeO}%Zgde?~5xHr>E3Pk{DlcC3qy?kO49djKx$E@%3tH$1PQ46i^Iz|D3Mipu*-V zi9u_rV8A$(R4|ZJJDGn+Va2ZrsAqyg%J?mhSSE{%zc1Dv2 z$|+BGmmh9zrb(WBv|n0Reufr|i3U|aIy}x&X4>qnPpu-6XMWvFU<+Bv+PA(`5Yzim z^w8j^e;*xKK&|s=s@e80Pp_E0*Wl>tFXmj-Gc8 zmDcHazV_ViY~-xj13bN%bo&*T)Q1vH28o~>KdW8a_jGz!j$+f!_ zZzY|TYMWFV5L`GnY`Yfwy#-2tW^G+{m57AmB=70_!qo%DsLm(GI4}xF9*E^+7qB58 z9=D!z63cNyE2+Jk086e80flNYU4_AIEb|;W+bUCR_jDQDy3Qz@wZo3;`C`0^EFzUc zHcP{iwQ{rtY=&{ev_D{RZrLe&3uWY~xooX23I1Ohe~Sdo3*CQ{nl$h=*mw|n-gYnM zNUk;F|M{2#FD#VEH8f^^XJ0^DgDQ##Ca2}S&FV*A{I7TVlJCiaOZl%bZGVMg*OIfA zYDUZjAWTtH_Jg^v%LgOFZ6AXl4XrLByouInM|2&7B`HWtJ_GncTA89nZSkg~cF$7! z+qBys!tP}XzR=yeMfngFx#H1wf+jG7KG+k!S2X##-*YRMJ>h*~X0l(r9BEEs{rv6o z(vI%`=8OJ~tc{g%IA_=FymUcmzdq@z)%@o7Db8`_6{(AnQiGFl!g!|%m?^C#W>x;D zjY!?q!iL)`8*2*_Xbi&wrdD z6;#-Aw41UP6ytW3s*n?=*CNg7nyfpM4B60u-~*5A`RSgt4pk^*KzsaSG8Q8uKZx0J z8>SDwTzh2l#rl1LegpZ9h3fvU$VnV`kD;<@6=E4r`OaiRjN;Yt9i-j zY36A?p2pDqW2N+g+iRkx$JH-e<+CmJr>02(MQf}PA6(Daq!A8*Va|bWgSu~umH%7S zE%(hmROYYyRjbzSc!TsxubNpAaKLSEd4QgAPf9EIdaYAlZ^FoZT~ogoyu}}_Hwpjk zi>;6@^$S-b@w6#VJJ9tM4hHMZ-PSJIDw{ir4h6)iNb&|`I9RgkcS4Zhw+vYZ3uutl zupWi@>swt$HsZeawmNJQhPXzk0DJq7zb67$u@_f#%NGT*CYN4vGf4pHV`p*Nz_6fa z{N?(MP%h1sFG8h$_VA+SI71$9iHA@fX#@6gtz}x_*KHK2tdxtDK%C28TG|vKPW+af zA?@ja5XI&oH}=}*kNm!EaZ@4cm?h>ik$A57*=nG-qt;y!(8y3z*+W(CRVVR8(XXZr zPitr0qMefs`c)ntJ`-tSJyZ^^aR1odtzh?1Gjy;irY__U;bQU=ZW3`qATCj zRKk3iRCYEkmtNIOo^x<5`&?eNUq1LP`jbz=&+RU+c38*&VQyg&bj9%G<%fj6MBY9n zB4&4fHARE2?Y?aTl3dNPh=Qp?Uquqh$87l9cNnLrm~n0M-p%)IIQ~I0HpxKiD7ZtZ zVhOCA`iu8*pY4@}1neDH3QDjwW|;ItlpP&^<#3Lf+K-{s{QTvp`|w{SUl{3#jno=Q z+ZNm#-Ibb~RlE|-Ba81L`^sO26Wpy7nfT!BEmG-YH%b}XtFivqauB*V8m+m-AeAY1 zv#}uL#}`t}gag~tp;uOOLT#6V>3AXRWw9Ni#JyGS4eOh<{))$kP(3c2Ue6mX2ID$3NbQ{K^%b+OYasARsr^kead1P6$VO<%AYb6#_A}*s6*V z&k>G%ga@R<1RB7FM+0%pYM3z=Mj%-nf@a4-Fv_kh#!D&zY%0e7FH=apC8& z13BS19FkTP$_rKY#&L30lF}~hI9l#)iyuCO?9>eelwiwPo^w(_x4Nc#!7gz$kWLPEo#e#1?u}x)hLGk z_BINgJY3bTz5881%RF0ARxpEhkJg&cu`o+UWxzn-CY$Z!SL4cv#8E!84X=b_`jQW~ zZ@Y>QfYMVPMl$?nF)TV^e?v0klIA97q>8?dq`>VyeKh7byWbt^ov;BL6Vw#?aX9Wx z!%fQ(1>@wm%UB)X8_S(6{YOdUTai!LoV8f(GXRJ!0>kbby_Y35YL8aU*j2T$Fc!A2 zj(C2^dglS2VHTNt{qAgF-Zwqn9#5=NppO^*GVuLomA^Y3dBj_fQo>$SwBWteWkgwI z?xfBXjkN_Fl2xMS7fkY3S9b|j&!BUJPx`(vG}KmOQrm0N?crkH!pre;KQb-!tBptr z(nuBONlMJ3l&fUHeL74+J5Dp{f>9V<5+#f>32=4_0>)~Pfx-(+WWX~$f9lcedEvn9 z&Z*-VRmlbS;~cxm+Hb!+%$xVXv+T-+!g$UGlmS2%YYHLnfZcqJV4w?|Ch(;`-Bm2j zey;vt&td1gaS%mF^50$-V!a$Rv?t%)<_8&^yd42}cF}#a4O%uq`#pv`-GLP7f(7rX z3b3_mZbir7O8HqDJH;2$UDKltE#?hEOBNOO4*Z2CMeeo3bDsO5b&a_b`)aSolK?Cmhn#c=oP@71N6uiM0v zCvOVtn^tH`u6;iw?yD1hAKtatYIF%fhMU?lQ4*65@Y(pp3|GG$6X$^tf00B@O7Kc0 zKbPr7st_TjPX%3Ml-P5qnO!vFSRp7brvhY6RhAU*bk}qJHz4#)_kHW_!k3z88Lc00 zJVcS(`UT+fOj+uh=`gmerPSxXUW(kc=G5*+LoBBvdq_U5Dht|(Ci+pIZ>WORCP}w^Xw;Pe)3^8?TQayYM3Vd145`Lu1|Ii>R%(I)D@?-%VIXVUHLK5E z)E`Zn0jMM5aNC&_U3rgFyQ!Zz>O0}d1C@OO-XQ}%?mkiWOm7BFwmL!Otn*QX3dlj1ubR4;ik9wlGMy9v|%U~`$1>Q2FR&)pqL^8 z8Ob?gKaW49sg4<$u~W!kWQmBzCW71=+qS>>y0S)(Y}5_bU^epUFbgS< zIwArr4n=}!m@^Zm#0;PUgjr1p2mNqDpGu8}GiOXTE#$Tm8g(&l~m$uM`w*$qD7!k9+0fLm$>8SGwBj z)`Nqfa9f_U=wzD4gBDYNMXodHx~thvE{-+6BqH#dm`liK(ngKEpRD3BC<)&g5Q$r+ ze)-E>k?}(ITa8ep8|bqlYsSQS!qgj+^IRR)Dw6xLH2F}&_{!Zx&lQsX#lFM4HGuq{95PkSGxZgPY^ zuD{})T@;xDzd4F=&^x$V3tYV5fBU!c`!$cmds5>)deRwr!W&j~OnJK2=r6HbqP zfntFF->K!i*Tr!&c7$9bK2q+)pm`8QqN3{$3{uu-1_X=%j-WU_%^pCb8tRs?y{xa? z-+|xF-tRqrD^h-m0jnXsTtmB(zMXc>bB#qOZ5WxqRsL;K@3xrF|L47AYyF3;G0$g3 zLdRgfi$Jf6Sf)YmSXPZctKEy$SGU!DdNXSa-kF(Qd8l>yhV@?bpbxHI8LoZ{EjhoR z5SMUNp!q4G`re7XExGHRw4_FxY)Fz68hn93 z*Nq@RcKs+T9u-F*Kq&l;P{Qbhsl>1SmUR!6Iy2$a*4P?S?r7JWUP5|KW8*;E5E5{` zB)I*qp3|m#?YM8b)fxl!Q_5VH^15DqZyt2QzMPVoMxGr%S?Z>FYKx{9-z&miocFCU zIb*`eRueQKyQ;8U|NTmZ*()B7^_M@r@5RN-XbV1u$}^CZ#V627_bieTC4Xq4Z?uS5 zMPkh;f9o0Xf1fk~c%D+f0b>7?nnfVv47Ouc;y-V9QPJwy5wVej`hDV=Uj}$XL{d?5 z1CD@*QjHrAJU3m9TO+W>f{GGz8rI;>d+I%{Z9*(MEXy}lxI1eF;FWAy-ya1;t9WR3 z7AuA3Mk=Td=IhS+Py7i5uSLn>D3L4fep6NBKJE2hAC}sqJ2IX&?fp5B4UJhpUb&Z( zr*pcMO-H*&uR%L963B9k*=uZF>}$b~%U)yc(C-VoSMpyUau9z1H0%`AKKXpH8z1}Z zGJ}=*p6(*CXGP>K6i@VwR(n5f`uNNTrtybId4si2>a-^ARCDhrOEKr9_&r;eF4;R_ zP6P5pB%J%btS6l=Bm+|pr)zdmp%)Sl0W*MfUD{Kj`>%_pFQ zuWZ_3_3v|-PG!Qbe*W}Kyyc&G_IUQDaZSqNK9ln%hY{3#?_@OHqA$x}79__!fi8(FNFTn#y9I*!?t7lP1idrOjEo|q6Ylj@iVS%A~;8lbG zxw!Zx+uZ3 zc}%Q1H#0u+z$Ka?6P}LGV@j^js^;ZOZ*J0ScYJYAepbw}cGOV)<5`Vr){Lq~zC*yO z`-I9~L9eEtr0%{EYNg+0JaeJP`5#|R!cuNw_>^eGM&_kdRI46wM--T_v*>>LEeZgk zqGI9`;)w^-A(2-2mOh)Y0mv|rKsZNWcLf)=NiLUB13%=&NZ9T3)L2@dh~~ z&Ap9p1I*rtR{G*)BQgsk9(HkT_=uQvVS7fom6I2NhQlLs@k3?P#k;}Da&@OY*w)0S z%JVJ1N@vvl+=t(W!>vprEbRr!bOz_*RVspiC15EhSOe{Ua{V&>8@ai(dtF};jFPB) zWb(#<7Vj&@i)2OKKlDi8)7Mw~vfcmRqqdKZ#!F7~OVyvVn<*z6UUVC|{QXdJ@UM!N zCAKU8>7mfG67Cb@U68z(IX%0Ot?J}2cv5E+ldXR@rT9Z|wy8_y&DnV;xo)}tn%}KO zYX{k}_jbhoS2At#ubt9R^|Y6{xV_M5Pvj_Yq^1u;WRZ#hVEe1D2p&;1Bib<0PDDTv z4OuI#OaB2YXnYSA2+n_gKfV8;me8nuFmpQ94TW#bw&IeMV~;pd7vLTn5Rh zWPsr@fU=p{z(=96+TB{o=p{e0%=;Db&X2h1l-r+~v7)u?zc?|LRviv6$f7~QplP^R zeA*q{xPzZbHEkV%<{l%@nKQjYo%sjZGYDq$97;JJDx;4GTe3g*Ond8EaQEHpq~Y)u zAz#bhTq!yynF5OAQXYQ$}|3)zL#qCwizw={In+Pflr9kI3z^u*DGEd~oBPoe=O zA>8-O245<2zH?~+<4nqFB8sGrkbq)SOcerA6=&3QYocX{+g{v2I%Gg4kefMIwy8^kE?e5I01~ebz$hBDD*&R&hgS#Tg^+u5}6ZL%%fmWJe~$6eU~#gS_{PstZc~B)eqn+&|NW3S^@S2BkSv0CA}4sWEP(F9du!HG zTPyM`&tjicn+UP}cD)>N1|=KMSY3^LCia~FJil%{U;n}`4zJMuf!$Ys@8&!Xo4j@T zFy%pQ?;pYC(_eS5E{^tp5t@oKe6sd{|AOrtlgC&|RX1qm-W_Oz$iNBF1272OjZI9e z_Jan5eLX_81`KCrkkuX zICB?=_?2bKaMFNutjQkghLlndHQXeDjYP$fDW7p*FyP@0I+`jF-1V6d2?mcG0NT=E ziTxJ*P>(?p#1XP>KHjH@sStRPYAv+{@kV|L`tv}v1GC}Nnz}wh&)UxlSR1#D*u1sq zt{2HFVbm8kd1B~YUbdE+rY`a1-I;BR7y6dltG%l_2-V)_Nt?gHFT*ZhvrP>Dsp$QE z{W(v6j(*S!|D}wGk+2s{`abv^?bdYmjSBW$x5f-ou|;~*ZS%y$KZ?eb z3L#wk+HT|Dq7!At6F!MJg(qghb6gt?zRV0U z`yL#Wlafb987j`eYrY8?RS#PUyU#tFv=6)Cb5PeJC*YDQd{>{3RC|%LZo=JxtIJX{ zzPctb!y)U8&sEf)R&_>OS>=y*IWpIBATt0pt6yPTW?fGEgp{?QA(EG;P=lHY73j zBHTsewv~KJDg8%fiees-${T#nJ(!t0HXRb`S?LVBUCJ-@^(YTtdysr|&H7W$q*o*H zX17&qjcoTQU1vn1Fx_IKD15Lx4Oc^!Jg4jQD$r|B%RM8OMog%Jqt2Im>|T!tnP*n- z^X^$ABDJMKSP|wxXG2tb`UWND9zkllEuJ^n*$SzFH7Ef;dQOppt@pdwW2L7^# z`Z;+jQ_2rfZG{TESm1z-K3W7OgN%$vUQW{b?JgC-61$cbgc zfq>&`o@Y~OOq?7-t@vUTy7B0t>DDE3Lnglw@rzG5M~or%kN51y$~ZfBuhrk<%Kw zK0&%iw=VS6f!Qi7jH;FI@{md8W)LO)IxieVh2r70i~N5sDmi!<#6LOn4{zlBVaEb$ z5+6?6$2KfTLPC#38m>LbTY6UWh;1rMs)O<>f>n($&wH}Hm0x`j#?!dnZDy3_wqrt$ zaO}iqzO6v#Jm==s(3Z5ij||;Wj#}*tB5T72PjZaVs{Su%zZvZqMt@@Vpfi)|TD>iJMbC zg&Cenu8#k}7+KOE&}7>X2e5Kflk4LL@GBzNbQg;HKr|b17E}({@7}gfl8Q(qvZj7- z--hOq@l?`8s=4pN>!i9p%(&FNW@ipf4C;$o#kO{z5esECj~FD*3y+a+;@=T6T}KOs zg;U@G;F}MYRxKypY(aeN04v+3AR9R!e8mV}qm7Q~sVv*J!xEo&Pk$%G~Ng z=T^M?qk;L1e9L=mSCj9=?yx>3&7-?2YIfa|UY&B&xVWHLL>e?V+0nInE;1$N8`9qP z=mQ*G$NaST>CfoxjmhB?^R0kE^y~%(!CAw7{|;=8F_)zgZ$Q5207#e>$BvBP=%kqY z3{85oq*#<+u<7y0&24}y;$}};?ReiVEhUF^40ARNjhAiK6uy=icJ+Fg^dp(}(_roW zOYok+@y0ECO9}(w{6Z-<*Fc?lW4YnKB8Qh~rrP$%jn!c)-N0-!6}d?QG(%eTxcE|}Jlqx$?dd>$bCg#3;1@8Yz%FJfn-9g|j*mQi_=#2&wo}J(1E=p2w+=r`x?VmfT z`P3#dCK5Z|?UUDjO^K^e5$msESL7H=?P88M38ns(Lu@4`g`3NFc-MGY#onNd(S$#>4}j$r6K z+KJK;g;yWxbrpdbe57{VoNzrI7G9Ot5?>sS-i)h1%*i&Z71eE7m=w|-8{QZ6ae6}!eY;!5I&g5v`$F|TA2@LFQFXq zZ&aw4gmhZtX$eE8fv$ps8vTMVDBQj~nZsCEc(C7{Gh)~!til(0eTZ%J{Dp{Zkxpj! zVhNVoib=8Mqg6kzCziY=Z+PUEO%x-PMwEG2^EbS9s!F<>KfX+pSn`$NPAFI_EhM8i z z-B~w8)qQV&h8TwK4(W~=Lb@A0cmbTL2`b> z_y6Y&Jm(dhbz<+e*Sgl{vh^{t$cQn+_rF_-+|p-e#o=Oz-w0NDRq{qawb(%7d1Xcw zMVcF8fI*yY;N9ZGJ)`&^dVlk?jcLayMe@BAA0rZJ<{Tq;73kKUQfRbCgO(njBTmGE$i$48)a;y2mEhCrsI$8VEX*gW}XM&;YGPZucb zjGtTc?8KAT3zZSxmozwpzWGf$F_bfyG93%S*iV`ykw{ZBzPha&QEXh;So7uPO8Usi zSX$sJuvYGa_y7l!6@3)MmLP6*uK;ILb3Z$$!*J#>K5Ll&La6>4PBN$~KROB~u*aI3 z*#_Yi$`riSrosB+Jiu_^0fvI>y3BdUyL{2hw_{STCL5dssG7{w+{LtKCtkbO=YJF3^T^v9 zHtu{g^4hsEm!yjW)k@o>M@i(QlGSbBOcY7X3Y|!gT z=io8MeRXDDjNTf{W1gQ(z`+VfMlH|)i*eR3= zJ6ZWS{7P57=)GE2z;K2(zBD=DZ7 zL6f+PQXkH1xzeHnoxG7OCpWP(G25LJG5|D~lJn+$tDjm0aGHrW;zo))e2LWG9f?6i)`ODq9@pUW zKfJG92^^XSzmj1%Nf)*U$v--&g(>AMTjdx>5Evoox8yw4q9RHcmq9w!_oH-6N}*}t z8$??o?bG)wCbl8_!`8|w@|Rd;gPjG#w0K=2vWwo7F$Z^ylW0c(#NQCD-$6s$uiw} zn+#JRoC^6V4M>*9DJkrawb@);$l&emeO^ojabx2u9F}~5-@L?0BbLFYd|3=N9v{4z zyYX3Ni)s~C-oh*B2&Gc1vPL+07LA-U?E<1?uvEeiNs9zqz5Zd9u*1P?*DSU( z&Tnnd$9&VcBjR4_FGCv(*K&lck+e|@Lp)Covw*2~9~?lU!!E-yTwVB-Vi|ZtGYm zZ< zqGj>)U*3ro+!cS*)l9$17t<7*{a+rg#YBVAM;9=S3Ej3FE^?`C`%Z86FZ_Pfe&Cq; zua%aLc11THI8LyCT0TVut401Xt^sqQd=ud-*pU=PZj(3Sv4v9>*E)#rPHdh zl?nfO3&T9iwD{EN8fl&dgvR9b<7ZBz!rvqPyBOZZ z4lpH)4zAp~T{BJ@Oivj--|L6_vOm(M)9fBi`&Wr|WNFse-MDh{5s9d`iw6iuu0Fp{ zV9ASgF4fExdTIGQBHx54@=e*#T90YVCFaBD>w5SZr86?~Mp4XG6i3Q$cy%^S_7I5q z0$sIc7u;4pR`uT-PaW<~uGb2=dcOJ5N41N_zzzBM!cGoA)<)ZlqqGpBs%vBy>y8m|E)hb)mmOHa_v56d?+6w(6*TBc4v1 zpIK_ESMxg=1=3fL*^kg9j_Ksb@UuyPb!5_$iJ1^b{rcoi_Ll8JuQ#jySL+rTTFu>^ z)YD!p8MVxL{CSG^RK`vc+raoKIX5y#z>_yl9e>`9ZJIsAA?T}GVs1$x)=T+@C@Q?K zj+^*8)B+R71}{ZkoG!mcHBjmla8YEd&&3$_0RQo1jjPmTW|ekukm*+0{`vh82!T<^ zh{}~3eUh41EBer17q7=pKhYjUADW-v&ZK}DjV-)fVfevLfHwCh`^Rxh!&z(Ev1c6> zDWh?-wO8vW9e^bjnTvSXFP4S=U|38jO*4Q%6c#vA;+3eORzgxB1(nqcW~z3r;q6cc zYvF4B{LIF7(|Q+un9R@-Pd!hbS|l!ZhBd4Mhas6xmD@>qP*41uFtjfcT)9EQ+Uu#z zf!b6+ogF1gruLHAdqFEC&6W;_PWnDriP5$lPi`dQa$P6aS<3y|Hy&EB@^3v|O(!Lw zGWsSLLUvJKWVznXr|>a6>zse(iNQsIo_h5T7oOAKn136|(;HL_Ozfmb_R{Gdm6J~~ zMp3fuTGw9vfx72C?Pl)|kMFu9n)=Z_Gse~PAI3r^D9F_j=Kr9w z-^79?vp%a+T2-WVlTyXW;2oTl7nC;V|Ip2j1t3H&mg~nZ^IPFS#vRriPO`HcqdqPG zg@&=X+-8Xfdg&h4!yUbYNzA$X_Ys^^cjISQx=e1h>>2x~H(9`3-A0QbCaF$0mkJe= z_ZL-s`)$1vj>UP>vXrAHKG}?&tOV9(3b+di*rzE$Q)i?MM*{{ zPv(?Op1%Up?k@C_!cqVMqu@q{kx>Lyx{s1Y7&!mtEqoRhbs8cOuj*i>VH5XWiKnqG zQ(Be%Ew`RoT~|p?#De4DNnaY*A_ev|Crig=k`kB&jn8*JtcOT_8%*71H#yKa7b`OR zy?|AA=jCkhE9J_0d%{RCnHt!Qmkd_G3lt$Xf~wRAkeRxWjSB?-aCVRWUf?n3NT$+1 z<*O2-RvW5t#z?8S2;MBj z0?App;W*N`!*nJZ(-}|PGC^%$bQ8RtnW;dg*?WVi3VUc8I{`LZKG%OX&}^{7uy{vZ zMza}E@KC6&Zk!#6pNwql5pcEz5s{~o!N}ng#SjRsN+O{CEaXE=U|ps#Wao}1f1m@I z2~umxbjf81w4#UeyMbzm%9(16rIURFnJ~TDgyBuC9E|2KhGtK=Bg2L_z-FSHjRM4(Y0<>rgF8KbcoxMRj@=o?5x3whK(ZIB(Wgy z(gWX1Ukns?WNdBQW1FPKuJ2qU!)JT?tl6^4Em7TRMuq3|Ty`TDt}m^vxlnaC)(8%) zFbPLCBM`A#``Elmc3d+r&>JVKJ)#Y#@VGzp>*U5RE&Kzh#GF!3#wIj*$z%9R%m`N{ zy+)OQ^ZV)DjXwXUi;$vaG`eXgHvMliH?f8nL5Etn4t%Qjm@n$l--Bjg`JsYRk`twk zTh_zcJ16|jJO(tV=Y?SfxloD*fYZ)KnS*y;cQcZ12uLcGC%FC48pPtN zE@JK8FUl+^RFg}#%2kb`XfH1{5;h;nmQotUh5-K)Ti;JKzt!-|GTDEjh_<`_6^F?y z2My9>&UAJHl`6d0OoY)yb|DT6A9bS13VurDaIexO0$EvQ4cbpK)}4JRqIBl$wNu@p zo5nm0r2Z&OzG^xALr{(+rfLZEFL3C;1ewU;Gkn>51UTC?M`6Euq5e75>lp@xCA4V7Go4KwpCt`!sbSt*7<|CQ$@!KC~7u)Kv=s)ZiOA4GAjr|beOx>z9wF2xY zEz@)2(y`b$Z@=^vA=#mWzfZVE0zqc328h!c4ye9to??ZT%3^LPYNHF(t3mVR!VX)3 z^rweaXP_%=M||pEiNTPhYtv9yjNnMnj9C8Jxo^FyFlEynce&Tj`-7D!QzNFQUoU%> z$>Kt+d6*00GQlS10fz(nE1`Db9F+t6x8bUuZUt5nd`?_EVJrKwozktp4_^!S&##Kw zocHH8a4B5Yc8CNzZxm^ITJ#4tW{9b3L?e)Yg~eS>Ot7dKwp$W%#5HxgZqc#Zk27!3 z9asM7(}9;6ke%B-NQvX^A-@W)@RiQPcQ_3#+Kts1!b`36n+Rr7`mg&$9INh#j%0-* zOXikjcr(WxXcO$QSxOtI>b`9}Mm`)_X7al*IG*if#A$K8#vj*B(oD>!u8bq9NSunu zJ%Uo4>ao*|@p@o&$83t&-gIis_B{2fD3pejv0gIIaJ#>vBIF-@`w*tb;mh!Rvd&ZT z+(2R0)fG%iOVeh0P%mqKlZjBWws z#UF+d|JmS!`Z9bbRKnPlj1SsC1Kmfo>(%cw{qZ5LSxEVHtj0icWoqkRo4OQE3TlZ# zyhtrosl)!PT~BOvK8Jq}_wI`xy^kB7QRgj1_+5JmX&--m{f6GFn~BxJ{d0)jfpp;v zb<$9~o_qkYcDWSa(utURH^rUHHB%&Gk>JzQh?AfB1_GrysyS~L*}*Q}r)yu?c^3!y ziSnEQPN?@O<&3lR?$X^-qxK_F7V0G~yHz$E3e_qmIW5O4>2<`(~ z1+A)2zWPzU6MNo9b42@ym0sRV|I~oFx<#BXIFswZawgL%vf&}F6{zkKX5G*nL0*fD zZ7KCA3#xWRt{U)+wdFRt-&1h?($;J(X_p>L@6s3v!<&3LqhH(?(_kGFFJR4>Pa+xG zH^N{>;ijUCY7;hsn$~k9#uQ}bY^GiCujYOsDYU=o`leDkv2Zfo?XCCGcz_ooSoywg zimaE#FzPs<`I7K8GMqx2TMtmF_#vJ|z6YHzwd! zFWVH`lUnnBu3w15TlfLn7fr~p$IhhOjp^%Q zSb2Q9GBGrRhJV5Og1lN`o2bX3G;_uQ<;trpPRgPZk82$~g$R>swWCr-S-P=0MoXsZ zwK9{O!fo`lZ^QyoL)@ZqouQH+rdbjaCr)X#UCxIzJ6}W`e1C-8zW!ZvcOKkm-iwmi zXt|ROWCdaY@L|c;{YvdDA?r|h{y;PLigJ~DHA&{_dSrIkTuerxDY&`bK%#4@utMgjPNl&O;|Qi0)h%V3hNM{p`$yry;jK-;nJ6d zMVA~RVqDJn5vhT=CGz}F6c_L}`dj6%wtVM4w6IZ$sw{^w*@VCU=5PBe)MlIC^buAx zkp?}2s+`ZiG^!js9t^}0VU*4O7Eq#h< zY8MfruR(s@OnF1ZI&|UQ<8OQatXKv|BBh#~rk@7&wW8y9K3hW@!LReDF>habPHG!7PO>~7(NVp(k8P0^b; z6Ta6J^8E)wyLsSbd~m%Ts*i?gZ7(`I4LZ2eZF9F)ibhr zT&A-6z3Y98l66B-lcO^w zE8m;S4xcQee;q2)eiqQ3Xt207L3@WkJf9}<08p&KN-)p|y`{J0d5Av!q!8$hF(N8@ z{^DK%DslEP;?wz^fUW0@LY`FyfI#p1?e9b~8?zsFTKwP01}{jLhBxo37Ohl}MXzYK zapIEpGsq}oV(}=5>dA8okR2GW5IIV|C#Z@6enAx&fnz{C?UI>j7u)@rOpg>|aT>fM z1|2Sb?wd?A=M?Tr+wxo!H4s|tzn0&5-Y18rGS`}(`>*6Hc;uZ*E)o#!f4Zt$ zyGR#bIVkm`=7rDKEA|PSmrr9@MHEz+i1xzFRn3)s7)3zPcXIkR7N-7#ziho#cQNMz zj8;v^BrVlEB!1X7eI{GbgBb^AcCQ2^cS21r%$sU|x42>FRdT zXroeTtJlhN>xtzeG>R{RWivI{r3Gu=UFDoRxV}{UcWnLXpL#9y-Q823e_h)M6gmNwgY|U7i zG3z$o`{wO`xoc)Mm){T{)_@-(a1D=Q={a6tAVUV#_$`gs3md+?P-XA9T z?u1x`2w{~B%r}PN>G5OdJv_t4oPl9iQYQ@o_Wo^~?7kP~Lb@FFkGGwh7xEUfzWbc6 z0j`z3^p^eTIK_upbLKrf09T|%2*olaW)e&*Cet#x1FrE^`C_g?8^=Ur(CIf05iJK# zZFf)Og!Y*-M>daU`(ynz^!nA^XUETKipD3C=1Qq5Y-~)dKi(L&3Z_&Dc7mk4E}caV zLFPFg^{?0qR#R3qf~HS}yeRv5w~~a>Bpd>CzI-YV`_2sH8-`sxsGdsVW+k^zM>qoR z!84?b`HJ+d8|ExdA^*eH#3oEvV{O%u|6FEe8awl8zTt3bYS|7xtgk`)QANs6Vo{RD zIUn1$hU}rh0-LV>QdOE;)qD1@-Z2|zOw)&AAI;tfP75_J|1d|%C;ivqgtFgo7=KGT z8^4XIdT4OecILr)eJ`OKNEg}Su4l_C1q@w8n#QB{1}d?>u6t9Ldr$H{ejHYbd*lV{ z2qM-h=_HAe#I*=EoV&#Wa-i7a>iOM8CsdXrO1sZ;yOStRC4wH}S`e%cN&tn#4x1v7 zIdn?6yu1*Z`B!v=(vhpEB^InVe4*OvPi8dPzE5mMlubGdz<8;k&L5lxiP-hbdFQ{y z#Bq;V9exTt@!$OC=s`8)knr*SVLDCF_Ln=iZ>N;)c$#}$K6tJYs|Qej)tD$nq$rJ4 zCdVJDBr)!Z<;oJ@7hAR5b<)kbsE=G0GVVOR6|Oxt#}uWbX*D$?cQ^h%G$Os|?xbQ{ z>x2w42`+nA=wd5m*zE zn7H;t1*$Ol&s0J*AGe-K&lHu0j%c#EU);wT0LE~c3{LJa3xc&sk(a_Xr1^a*pH8z7 z>PH$b0cGa19+}`2+aKk80?gr!m{5x-*SHgX!*+>dUH-u?#xMcmgxt!-@+H7=bZ2Tr zzyeCWRXgWg@jM#dY8^itk8g#SABCnaTz$^BfKC7U{E2Dz*Uam8HXd9%zvQ2Jjecpk zYxcGa&^R=~#v|pU>T=r?(aIcgdP<SdBa+zcMLy(JJ1~puW1oR4PvER zG@7|LUb;NWdK~QhBeicr`_Rfl)lUBaFZ~d~F`*vPfET}LSD&5&KlF6PmFCUwruJa1rkG_$InZKUe8+$pO z$9k=3#-y%g*{bH8@^EHckrKOygC>$^M5O|npQZ7o9)Rq`e-Ec($J6?S&RWemcsB>lTmRcVKsI8k|ny=nW3Q}Hq zj=iAJ_Ljq6PPZyPv3zToh^T`_)Nyms&X8=7Eee38gdg{Q_NG=OtzC<2GU`C?q;1bq z^8{F(#(gLdPT!t%JK~`)>xZcJe8kTrF2JP(M+kZs5|h6-jh>El63AprN-p=<{$TWl zKBD4Z)1MFZ(;qA8)Qx1EA}nJjR&j6+D6j}hUxl1JeGNgQ?a;Ry;t6TrjrmD$gy__4 zqpV^cV&9dGz6b-Mbf*`#p*iCcmAtqSj#-IoBuwLD@pu4bHdZ6LbJoaxCk)=_4=bl( z*8xcu9E6ACiRHHd9wotk+@y!ED=a}TO0vZ-#JV+2w%~9H(s2Sv`H!$kgc2dPAX$G) zuB?Qf3t{K${9XYj;Nk#-r-;gMbrr-E=?1{!5wg0^`p( z4!C^#1>=Ate|(&wo5bATjI`J{-Vi*0l<~vEh6QdF=V)nG=Oil<*_}sp5e`qWz2zjMyQOB~)}G zQJayLY&>(%RZ4*7#xuq7uN{0jIrJhN_TN>c0|0AE_@e-OC3_XXAb>Qhgj$ioxFl%* zorOPs&B@0F{l&dC<%a%dH3`1LTE=*Y=VueFfK5q;1>3=BhosMpp16w7c|0l;6LP+1~T zfiG+R1|(sHgt;nZbn5Jv=au!b5?=crn|ju5h252A&68$VkD_&D2!`rl%6jTz!cYjL Y?*G>deNH!g09?VM0wMu`|3{Gj2W%$_b^rhX literal 0 HcmV?d00001 diff --git a/tests/CMakeLists.txt b/tests/CMakeLists.txt index 5366f848b09..295bc48cf53 100644 --- a/tests/CMakeLists.txt +++ b/tests/CMakeLists.txt @@ -74,3 +74,14 @@ add_test(NAME ${TEST_TARGET} -m ${PROJECT_SOURCE_DIR}/models/for-tests-ggml-large.bin -f ${PROJECT_SOURCE_DIR}/samples/jfk.wav) set_tests_properties(${TEST_TARGET} PROPERTIES LABELS "large") + +if (WHISPER_FFMPEG) + set(TEST_TARGET test-main-tiny-mp3) + # Check with reviewers: any way to check the output transcription via ctest (diff, ...)? + add_test(NAME ${TEST_TARGET} + COMMAND $ + -m ${PROJECT_SOURCE_DIR}/models/for-tests-ggml-tiny.en.bin + -f ${PROJECT_SOURCE_DIR}/samples/jfk.mp3) + set_tests_properties(${TEST_TARGET} PROPERTIES LABELS "tiny;mp3") +endif() + From c10db6ea2883a4f77440fa8caeb296a0e351a58c Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Tue, 21 May 2024 18:44:37 +0300 Subject: [PATCH 13/17] release : v1.6.1 --- CMakeLists.txt | 2 +- bindings/ios | 2 +- bindings/javascript/package.json | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index 3eb12c10783..541be8a5d57 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -3,7 +3,7 @@ cmake_minimum_required (VERSION 3.5) # Allow for the creation of solution folders. set_property(GLOBAL PROPERTY USE_FOLDERS ON) -project(whisper.cpp VERSION 1.6.0) +project(whisper.cpp VERSION 1.6.1) set(SOVERSION 1) # Add path to modules diff --git a/bindings/ios b/bindings/ios index 5cfcfb0801b..9a32de38144 160000 --- a/bindings/ios +++ b/bindings/ios @@ -1 +1 @@ -Subproject commit 5cfcfb0801be756d8347822b472e4b5e343f403f +Subproject commit 9a32de3814477ad2e598d4a550fcab4b23a9c576 diff --git a/bindings/javascript/package.json b/bindings/javascript/package.json index 354d0ce903c..da6a9efdc6c 100644 --- a/bindings/javascript/package.json +++ b/bindings/javascript/package.json @@ -1,6 +1,6 @@ { "name": "whisper.cpp", - "version": "1.6.0", + "version": "1.6.1", "description": "Whisper speech recognition", "main": "whisper.js", "scripts": { From 22d46b7ba4620e2db1281e210d0186863cffcec0 Mon Sep 17 00:00:00 2001 From: Todd Date: Wed, 22 May 2024 16:02:52 -0400 Subject: [PATCH 14/17] ruby : update bindings (#2154) * update library files * update whispercpp * not needed for gem --- bindings/ruby/Rakefile | 12 + bindings/ruby/ext/ggml-backend-impl.h | 118 +- bindings/ruby/ext/ggml-backend.c | 2205 ++++- bindings/ruby/ext/ggml-backend.h | 217 +- bindings/ruby/ext/ggml-common.h | 1853 ++++ bindings/ruby/ext/ggml-cuda.h | 43 + bindings/ruby/ext/ggml-impl.h | 51 +- bindings/ruby/ext/ggml-kompute.h | 46 + bindings/ruby/ext/ggml-metal.h | 66 + bindings/ruby/ext/ggml-opencl.h | 36 + bindings/ruby/ext/ggml-quants.c | 12622 +++++++++++++++++------- bindings/ruby/ext/ggml-quants.h | 333 +- bindings/ruby/ext/ggml-sycl.h | 49 + bindings/ruby/ext/ggml-vulkan.h | 29 + bindings/ruby/whispercpp.gemspec | 28 + 15 files changed, 13247 insertions(+), 4461 deletions(-) create mode 100644 bindings/ruby/Rakefile create mode 100644 bindings/ruby/ext/ggml-common.h create mode 100644 bindings/ruby/ext/ggml-cuda.h create mode 100644 bindings/ruby/ext/ggml-kompute.h create mode 100644 bindings/ruby/ext/ggml-metal.h create mode 100644 bindings/ruby/ext/ggml-opencl.h create mode 100644 bindings/ruby/ext/ggml-sycl.h create mode 100644 bindings/ruby/ext/ggml-vulkan.h create mode 100644 bindings/ruby/whispercpp.gemspec diff --git a/bindings/ruby/Rakefile b/bindings/ruby/Rakefile new file mode 100644 index 00000000000..354d8ef2547 --- /dev/null +++ b/bindings/ruby/Rakefile @@ -0,0 +1,12 @@ +require 'rake/clean' + require 'rubygems/package' + +desc 'Build gem' +task :package do + spec_source = File.read File.join(File.dirname(__FILE__),'whispercpp.gemspec') + spec = nil + # see: http://gist.github.com/16215 + Thread.new { spec = eval("#{spec_source}") }.join + spec.validate + Gem::Package.build(spec) +end diff --git a/bindings/ruby/ext/ggml-backend-impl.h b/bindings/ruby/ext/ggml-backend-impl.h index 31788cd6baa..f121e1de420 100644 --- a/bindings/ruby/ext/ggml-backend-impl.h +++ b/bindings/ruby/ext/ggml-backend-impl.h @@ -12,31 +12,63 @@ extern "C" { // Backend buffer // + // buffer type + typedef void * ggml_backend_buffer_type_context_t; + + struct ggml_backend_buffer_type_i { + const char * (*GGML_CALL get_name) (ggml_backend_buffer_type_t buft); + ggml_backend_buffer_t (*GGML_CALL alloc_buffer) (ggml_backend_buffer_type_t buft, size_t size); + size_t (*GGML_CALL get_alignment) (ggml_backend_buffer_type_t buft); // tensor alignment + size_t (*GGML_CALL get_max_size) (ggml_backend_buffer_type_t buft); // allocation max size + size_t (*GGML_CALL get_alloc_size) (ggml_backend_buffer_type_t buft, const struct ggml_tensor * tensor); // data size needed to allocate the tensor, including padding + bool (*GGML_CALL supports_backend)(ggml_backend_buffer_type_t buft, ggml_backend_t backend); // check if the buffer type is usable by the backend + // check if tensor data is in host memory + // should be equivalent to supports_backend(buft, ggml_backend_cpu_init()) + bool (*GGML_CALL is_host) (ggml_backend_buffer_type_t buft); + }; + + struct ggml_backend_buffer_type { + struct ggml_backend_buffer_type_i iface; + ggml_backend_buffer_type_context_t context; + }; + + // buffer typedef void * ggml_backend_buffer_context_t; struct ggml_backend_buffer_i { - void (*free_buffer) (ggml_backend_buffer_t buffer); - void * (*get_base) (ggml_backend_buffer_t buffer); // get base pointer - size_t (*get_alloc_size)(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); // pre-allocation callback - void (*init_tensor) (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); // post-allocation callback - void (*free_tensor) (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); // pre-free callback + const char * (*GGML_CALL get_name) (ggml_backend_buffer_t buffer); + void (*GGML_CALL free_buffer)(ggml_backend_buffer_t buffer); + void * (*GGML_CALL get_base) (ggml_backend_buffer_t buffer); + void (*GGML_CALL init_tensor)(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); + void (*GGML_CALL set_tensor) (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); + void (*GGML_CALL get_tensor) (ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); + bool (*GGML_CALL cpy_tensor) (ggml_backend_buffer_t buffer, const struct ggml_tensor * src, struct ggml_tensor * dst); // dst is in the buffer, src may be in any buffer + void (*GGML_CALL clear) (ggml_backend_buffer_t buffer, uint8_t value); + void (*GGML_CALL reset) (ggml_backend_buffer_t buffer); // reset any internal state due to tensor initialization, such as tensor extras }; struct ggml_backend_buffer { - struct ggml_backend_buffer_i iface; - - ggml_backend_t backend; + struct ggml_backend_buffer_i iface; + ggml_backend_buffer_type_t buft; ggml_backend_buffer_context_t context; - size_t size; + enum ggml_backend_buffer_usage usage; }; - GGML_API ggml_backend_buffer_t ggml_backend_buffer_init( - struct ggml_backend * backend, + GGML_CALL ggml_backend_buffer_t ggml_backend_buffer_init( + ggml_backend_buffer_type_t buft, struct ggml_backend_buffer_i iface, ggml_backend_buffer_context_t context, size_t size); + // do not use directly, use ggml_backend_tensor_copy instead + bool ggml_backend_buffer_copy_tensor(const struct ggml_tensor * src, struct ggml_tensor * dst); + + // buffer that contains a collection of buffers + GGML_CALL ggml_backend_buffer_t ggml_backend_multi_buffer_alloc_buffer(ggml_backend_buffer_t * buffers, size_t n_buffers); + GGML_CALL bool ggml_backend_buffer_is_multi_buffer(ggml_backend_buffer_t buffer); + GGML_CALL void ggml_backend_multi_buffer_set_usage(ggml_backend_buffer_t buffer, enum ggml_backend_buffer_usage usage); + // // Backend // @@ -44,44 +76,66 @@ extern "C" { typedef void * ggml_backend_context_t; struct ggml_backend_i { - const char * (*get_name)(ggml_backend_t backend); + const char * (*GGML_CALL get_name)(ggml_backend_t backend); - void (*free)(ggml_backend_t backend); + void (*GGML_CALL free)(ggml_backend_t backend); // buffer allocation - ggml_backend_buffer_t (*alloc_buffer)(ggml_backend_t backend, size_t size); + ggml_backend_buffer_type_t (*GGML_CALL get_default_buffer_type)(ggml_backend_t backend); - // get buffer alignment - size_t (*get_alignment)(ggml_backend_t backend); + // (optional) asynchronous tensor data access + void (*GGML_CALL set_tensor_async)(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); + void (*GGML_CALL get_tensor_async)(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); + bool (*GGML_CALL cpy_tensor_async)(ggml_backend_t backend_src, ggml_backend_t backend_dst, const struct ggml_tensor * src, struct ggml_tensor * dst); - // tensor data access - // these functions can be asynchronous, helper functions are provided for synchronous access that automatically call synchronize - void (*set_tensor_async)(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); - void (*get_tensor_async)(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); - void (*synchronize) (ggml_backend_t backend); + // (optional) complete all pending operations + void (*GGML_CALL synchronize)(ggml_backend_t backend); - // (optional) copy tensor between different backends, allow for single-copy tranfers - void (*cpy_tensor_from)(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst); - void (*cpy_tensor_to) (ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst); + // compute graph with a plan (not used currently) + ggml_backend_graph_plan_t (*GGML_CALL graph_plan_create) (ggml_backend_t backend, const struct ggml_cgraph * cgraph); + void (*GGML_CALL graph_plan_free) (ggml_backend_t backend, ggml_backend_graph_plan_t plan); // compute graph with a plan - ggml_backend_graph_plan_t (*graph_plan_create) (ggml_backend_t backend, struct ggml_cgraph * cgraph); - void (*graph_plan_free) (ggml_backend_t backend, ggml_backend_graph_plan_t plan); - void (*graph_plan_compute)(ggml_backend_t backend, ggml_backend_graph_plan_t plan); - - // compute graph without a plan - bool (*graph_compute)(ggml_backend_t backend, struct ggml_cgraph * cgraph); + enum ggml_status (*GGML_CALL graph_plan_compute)(ggml_backend_t backend, ggml_backend_graph_plan_t plan); + // compute graph without a plan (async) + enum ggml_status (*GGML_CALL graph_compute) (ggml_backend_t backend, struct ggml_cgraph * cgraph); // check if the backend supports an operation - bool (*supports_op)(ggml_backend_t backend, const struct ggml_tensor * op); + bool (*GGML_CALL supports_op)(ggml_backend_t backend, const struct ggml_tensor * op); + + // check if the backend wants to run an operation, even if the weights are allocated in a CPU buffer + // these should be expensive operations with large batch sizes that may benefit from running on this backend + // even if the weight has to be copied from the CPU temporarily + bool (*GGML_CALL offload_op)(ggml_backend_t backend, const struct ggml_tensor * op); + + // (optional) event synchronization + ggml_backend_event_t (*GGML_CALL event_new) (ggml_backend_t backend); + void (*GGML_CALL event_free) (ggml_backend_event_t event); + void (*GGML_CALL event_record) (ggml_backend_event_t event); + void (*GGML_CALL event_wait) (ggml_backend_t backend, ggml_backend_event_t event); + void (*GGML_CALL event_synchronize) (ggml_backend_event_t event); }; struct ggml_backend { - struct ggml_backend_i iface; + ggml_guid_t guid; + struct ggml_backend_i iface; ggml_backend_context_t context; }; + struct ggml_backend_event { + ggml_backend_t backend; + void * context; + }; + + // + // Backend registry + // + + typedef ggml_backend_t (*GGML_CALL ggml_backend_init_fn)(const char * params, void * user_data); + + GGML_CALL void ggml_backend_register(const char * name, ggml_backend_init_fn init_fn, ggml_backend_buffer_type_t default_buffer_type, void * user_data); + #ifdef __cplusplus } #endif diff --git a/bindings/ruby/ext/ggml-backend.c b/bindings/ruby/ext/ggml-backend.c index 128e33ce630..402d86ef3ac 100644 --- a/bindings/ruby/ext/ggml-backend.c +++ b/bindings/ruby/ext/ggml-backend.c @@ -9,31 +9,76 @@ #include #include -#define UNUSED GGML_UNUSED #define MAX(a, b) ((a) > (b) ? (a) : (b)) +// backend buffer type + +const char * ggml_backend_buft_name(ggml_backend_buffer_type_t buft) { + return buft->iface.get_name(buft); +} + +GGML_CALL ggml_backend_buffer_t ggml_backend_buft_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + return buft->iface.alloc_buffer(buft, size); +} + +size_t ggml_backend_buft_get_alignment(ggml_backend_buffer_type_t buft) { + return buft->iface.get_alignment(buft); +} + +size_t ggml_backend_buft_get_max_size(ggml_backend_buffer_type_t buft) { + // get_max_size is optional, defaults to SIZE_MAX + if (buft->iface.get_max_size) { + return buft->iface.get_max_size(buft); + } + return SIZE_MAX; +} + +GGML_CALL size_t ggml_backend_buft_get_alloc_size(ggml_backend_buffer_type_t buft, struct ggml_tensor * tensor) { + // get_alloc_size is optional, defaults to ggml_nbytes + if (buft->iface.get_alloc_size) { + size_t size = buft->iface.get_alloc_size(buft, tensor); + assert(size >= ggml_nbytes(tensor)); + return size; + } + return ggml_nbytes(tensor); +} + +bool ggml_backend_buft_supports_backend(ggml_backend_buffer_type_t buft, ggml_backend_t backend) { + return buft->iface.supports_backend(buft, backend); +} + +bool ggml_backend_buft_is_host(ggml_backend_buffer_type_t buft) { + if (buft->iface.is_host) { + return buft->iface.is_host(buft); + } + return false; +} + // backend buffer -ggml_backend_buffer_t ggml_backend_buffer_init( - struct ggml_backend * backend, +GGML_CALL ggml_backend_buffer_t ggml_backend_buffer_init( + ggml_backend_buffer_type_t buft, struct ggml_backend_buffer_i iface, ggml_backend_buffer_context_t context, size_t size) { ggml_backend_buffer_t buffer = malloc(sizeof(struct ggml_backend_buffer)); - GGML_ASSERT(iface.get_base != NULL); - (*buffer) = (struct ggml_backend_buffer) { /* .interface = */ iface, - /* .backend = */ backend, + /* .buft = */ buft, /* .context = */ context, /* .size = */ size, + /* .usage = */ GGML_BACKEND_BUFFER_USAGE_ANY }; return buffer; } +const char * ggml_backend_buffer_name(ggml_backend_buffer_t buffer) { + return buffer->iface.get_name(buffer); +} + void ggml_backend_buffer_free(ggml_backend_buffer_t buffer) { if (buffer == NULL) { return; @@ -45,10 +90,6 @@ void ggml_backend_buffer_free(ggml_backend_buffer_t buffer) { free(buffer); } -size_t ggml_backend_buffer_get_alignment(ggml_backend_buffer_t buffer) { - return ggml_backend_get_alignment(buffer->backend); -} - size_t ggml_backend_buffer_get_size(ggml_backend_buffer_t buffer) { return buffer->size; } @@ -61,32 +102,67 @@ void * ggml_backend_buffer_get_base(ggml_backend_buffer_t buffer) { return base; } +GGML_CALL void ggml_backend_buffer_init_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { + // init_tensor is optional + if (buffer->iface.init_tensor) { + buffer->iface.init_tensor(buffer, tensor); + } +} + +size_t ggml_backend_buffer_get_alignment (ggml_backend_buffer_t buffer) { + return ggml_backend_buft_get_alignment(ggml_backend_buffer_get_type(buffer)); +} + +size_t ggml_backend_buffer_get_max_size(ggml_backend_buffer_t buffer) { + return ggml_backend_buft_get_max_size(ggml_backend_buffer_get_type(buffer)); +} + size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { - // get_alloc_size is optional, defaults to ggml_nbytes - if (buffer->iface.get_alloc_size) { - return buffer->iface.get_alloc_size(buffer, tensor); + return ggml_backend_buft_get_alloc_size(ggml_backend_buffer_get_type(buffer), tensor); +} + +void ggml_backend_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { + buffer->iface.clear(buffer, value); +} + +bool ggml_backend_buffer_is_host(ggml_backend_buffer_t buffer) { + return ggml_backend_buft_is_host(ggml_backend_buffer_get_type(buffer)); +} + +void ggml_backend_buffer_set_usage(ggml_backend_buffer_t buffer, enum ggml_backend_buffer_usage usage) { + buffer->usage = usage; + + // FIXME: add a generic callback to the buffer interface + if (ggml_backend_buffer_is_multi_buffer(buffer)) { + ggml_backend_multi_buffer_set_usage(buffer, usage); } - return ggml_nbytes(tensor); } -void ggml_backend_buffer_init_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { - // init_tensor is optional - if (buffer->iface.init_tensor) { - buffer->iface.init_tensor(buffer, tensor); +ggml_backend_buffer_type_t ggml_backend_buffer_get_type(ggml_backend_buffer_t buffer) { + return buffer->buft; +} + +void ggml_backend_buffer_reset(ggml_backend_buffer_t buffer) { + if (buffer->iface.reset) { + buffer->iface.reset(buffer); } } -void ggml_backend_buffer_free_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { - // free_tensor is optional - if (buffer->iface.free_tensor) { - buffer->iface.free_tensor(buffer, tensor); +bool ggml_backend_buffer_copy_tensor(const struct ggml_tensor * src, struct ggml_tensor * dst) { + ggml_backend_buffer_t dst_buf = dst->view_src ? dst->view_src->buffer : dst->buffer; + if (dst_buf->iface.cpy_tensor) { + return src->buffer->iface.cpy_tensor(dst_buf, src, dst); } + return false; } // backend -ggml_backend_t ggml_get_backend(const struct ggml_tensor * tensor) { - return tensor->buffer ? tensor->buffer->backend : NULL; +ggml_guid_t ggml_backend_guid(ggml_backend_t backend) { + if (backend == NULL) { + return NULL; + } + return backend->guid; } const char * ggml_backend_name(ggml_backend_t backend) { @@ -104,59 +180,105 @@ void ggml_backend_free(ggml_backend_t backend) { backend->iface.free(backend); } +ggml_backend_buffer_type_t ggml_backend_get_default_buffer_type(ggml_backend_t backend) { + return backend->iface.get_default_buffer_type(backend); +} + ggml_backend_buffer_t ggml_backend_alloc_buffer(ggml_backend_t backend, size_t size) { - return backend->iface.alloc_buffer(backend, size); + return ggml_backend_buft_alloc_buffer(ggml_backend_get_default_buffer_type(backend), size); } size_t ggml_backend_get_alignment(ggml_backend_t backend) { - return backend->iface.get_alignment(backend); + return ggml_backend_buft_get_alignment(ggml_backend_get_default_buffer_type(backend)); +} + +size_t ggml_backend_get_max_size(ggml_backend_t backend) { + return ggml_backend_buft_get_max_size(ggml_backend_get_default_buffer_type(backend)); } -void ggml_backend_tensor_set_async(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - ggml_get_backend(tensor)->iface.set_tensor_async(ggml_get_backend(tensor), tensor, data, offset, size); +void ggml_backend_tensor_set_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + GGML_ASSERT(tensor->data != NULL && "tensor not allocated"); + GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor write out of bounds"); + + if (backend->iface.set_tensor_async == NULL) { + ggml_backend_tensor_set(tensor, data, offset, size); + } else { + backend->iface.set_tensor_async(backend, tensor, data, offset, size); + } } -void ggml_backend_tensor_get_async(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - ggml_get_backend(tensor)->iface.get_tensor_async(ggml_get_backend(tensor), tensor, data, offset, size); +void ggml_backend_tensor_get_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + GGML_ASSERT(tensor->data != NULL && "tensor not allocated"); + GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor read out of bounds"); + + if (backend->iface.get_tensor_async == NULL) { + ggml_backend_tensor_get(tensor, data, offset, size); + } else { + backend->iface.get_tensor_async(backend, tensor, data, offset, size); + } } -void ggml_backend_tensor_set(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - ggml_backend_t backend = ggml_get_backend(tensor); +GGML_CALL void ggml_backend_tensor_set(struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + ggml_backend_buffer_t buf = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; + GGML_ASSERT(buf != NULL && "tensor buffer not set"); GGML_ASSERT(tensor->data != NULL && "tensor not allocated"); - GGML_ASSERT(backend != NULL && "tensor backend not set"); + GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor write out of bounds"); - backend->iface.set_tensor_async(backend, tensor, data, offset, size); - backend->iface.synchronize(backend); + if (!size) { + return; + } + + buf->iface.set_tensor(buf, tensor, data, offset, size); } -void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - ggml_backend_t backend = ggml_get_backend(tensor); +GGML_CALL void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + ggml_backend_buffer_t buf = tensor->view_src ? tensor->view_src->buffer : tensor->buffer; + GGML_ASSERT(buf != NULL && "tensor buffer not set"); GGML_ASSERT(tensor->data != NULL && "tensor not allocated"); - GGML_ASSERT(backend != NULL && "tensor backend not set"); + GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor read out of bounds"); - backend->iface.get_tensor_async(backend, tensor, data, offset, size); - backend->iface.synchronize(backend); + if (!size) { + return; + } + + buf->iface.get_tensor(buf, tensor, data, offset, size); } void ggml_backend_synchronize(ggml_backend_t backend) { + if (backend->iface.synchronize == NULL) { + return; + } + backend->iface.synchronize(backend); } ggml_backend_graph_plan_t ggml_backend_graph_plan_create(ggml_backend_t backend, struct ggml_cgraph * cgraph) { + GGML_ASSERT(backend->iface.graph_plan_create != NULL); + return backend->iface.graph_plan_create(backend, cgraph); } void ggml_backend_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { + GGML_ASSERT(backend->iface.graph_plan_free != NULL); + backend->iface.graph_plan_free(backend, plan); } -void ggml_backend_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { - backend->iface.graph_plan_compute(backend, plan); +enum ggml_status ggml_backend_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { + GGML_ASSERT(backend->iface.graph_plan_compute != NULL); + + return backend->iface.graph_plan_compute(backend, plan); +} + +enum ggml_status ggml_backend_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { + enum ggml_status err = ggml_backend_graph_compute_async(backend, cgraph); + ggml_backend_synchronize(backend); + return err; } -bool ggml_backend_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { +enum ggml_status ggml_backend_graph_compute_async(ggml_backend_t backend, struct ggml_cgraph * cgraph) { return backend->iface.graph_compute(backend, cgraph); } @@ -164,6 +286,13 @@ bool ggml_backend_supports_op(ggml_backend_t backend, const struct ggml_tensor * return backend->iface.supports_op(backend, op); } +bool ggml_backend_offload_op(ggml_backend_t backend, const struct ggml_tensor * op) { + if (backend->iface.offload_op != NULL) { + return backend->iface.offload_op(backend, op); + } + return false; +} + // backend copy static bool ggml_are_same_layout(const struct ggml_tensor * a, const struct ggml_tensor * b) { @@ -182,27 +311,20 @@ static bool ggml_are_same_layout(const struct ggml_tensor * a, const struct ggml } void ggml_backend_tensor_copy(struct ggml_tensor * src, struct ggml_tensor * dst) { - //printf("src: %s ne: [%d %d %d %d] nb: [%d %d %d %d]\n", src->name, (int)src->ne[0], (int)src->ne[1], (int)src->ne[2], (int)src->ne[3], (int)src->nb[0], (int)src->nb[1], (int)src->nb[2], (int)src->nb[3]); - //printf("dst: %s ne: [%d %d %d %d] nb: [%d %d %d %d]\n", dst->name, (int)dst->ne[0], (int)dst->ne[1], (int)dst->ne[2], (int)dst->ne[3], (int)dst->nb[0], (int)dst->nb[1], (int)dst->nb[2], (int)dst->nb[3]); GGML_ASSERT(ggml_are_same_layout(src, dst) && "cannot copy tensors with different layouts"); - // fprintf(stderr, "cpy tensor %s from %s to %s (%lu bytes)\n", src->name, ggml_backend_name(src->backend), ggml_backend_name(dst->backend), ggml_nbytes(src)); - if (src == dst) { return; } - // TODO: allow backends to support copy to/from same backend - - if (ggml_get_backend(dst)->iface.cpy_tensor_from != NULL) { - ggml_get_backend(dst)->iface.cpy_tensor_from(ggml_get_backend(dst)->context, src, dst); - } else if (ggml_get_backend(src)->iface.cpy_tensor_to != NULL) { - ggml_get_backend(src)->iface.cpy_tensor_to(ggml_get_backend(src)->context, src, dst); - } else { - // shouldn't be hit when copying from/to CPU - #ifndef NDEBUG - fprintf(stderr, "ggml_backend_tensor_copy: neither cpy_tensor_from nor cpy_tensor_to are implemented for backends %s and %s, falling back to get/set\n", ggml_backend_name(src->buffer->backend), ggml_backend_name(dst->buffer->backend)); - #endif + if (ggml_backend_buffer_is_host(src->buffer)) { + ggml_backend_tensor_set(dst, src->data, 0, ggml_nbytes(src)); + } else if (ggml_backend_buffer_is_host(dst->buffer)) { + ggml_backend_tensor_get(src, dst->data, 0, ggml_nbytes(src)); + } else if (!ggml_backend_buffer_copy_tensor(src, dst)) { +#ifndef NDEBUG + fprintf(stderr, "%s: warning: slow copy from %s to %s\n", __func__, ggml_backend_buffer_name(src->buffer), ggml_backend_buffer_name(dst->buffer)); +#endif size_t nbytes = ggml_nbytes(src); void * data = malloc(nbytes); ggml_backend_tensor_get(src, data, 0, nbytes); @@ -211,318 +333,846 @@ void ggml_backend_tensor_copy(struct ggml_tensor * src, struct ggml_tensor * dst } } -// backend CPU +void ggml_backend_tensor_copy_async(ggml_backend_t backend_src, ggml_backend_t backend_dst, struct ggml_tensor * src, struct ggml_tensor * dst) { + GGML_ASSERT(ggml_are_same_layout(src, dst) && "cannot copy tensors with different layouts"); -struct ggml_backend_cpu_context { - int n_threads; - void * work_data; - size_t work_size; -}; + if (src == dst) { + return; + } -static const char * ggml_backend_cpu_name(ggml_backend_t backend) { - return "CPU"; + if (backend_dst->iface.cpy_tensor_async != NULL) { + if (backend_dst->iface.cpy_tensor_async(backend_src, backend_dst, src, dst)) { + return; + } + } - UNUSED(backend); + // an async copy would normally happen after all the queued operations on both backends are completed + // sync src, set_async dst + if (ggml_backend_buffer_is_host(src->buffer)) { + ggml_backend_synchronize(backend_src); + ggml_backend_tensor_set_async(backend_dst, dst, src->data, 0, ggml_nbytes(src)); + } else { + ggml_backend_synchronize(backend_src); + ggml_backend_tensor_copy(src, dst); + ggml_backend_synchronize(backend_dst); + } } -static void ggml_backend_cpu_free(ggml_backend_t backend) { - struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; - free(cpu_ctx->work_data); - free(cpu_ctx); - free(backend); +// events + +ggml_backend_event_t ggml_backend_event_new(ggml_backend_t backend) { + if (backend->iface.event_new == NULL) { + return NULL; + } + return backend->iface.event_new(backend); +} + +void ggml_backend_event_free(ggml_backend_event_t event) { + if (event == NULL) { + return; + } + event->backend->iface.event_free(event); } -static void * ggml_backend_cpu_buffer_get_base(ggml_backend_buffer_t buffer) { - return (void *)buffer->context; +void ggml_backend_event_record(ggml_backend_event_t event) { + GGML_ASSERT(event->backend->iface.event_record != NULL); + + event->backend->iface.event_record(event); } -static void ggml_backend_cpu_buffer_free_buffer(ggml_backend_buffer_t buffer) { - free(buffer->context); - UNUSED(buffer); +void ggml_backend_event_synchronize(ggml_backend_event_t event) { + GGML_ASSERT(event->backend->iface.event_synchronize != NULL); + + event->backend->iface.event_synchronize(event); } -static struct ggml_backend_buffer_i cpu_backend_buffer_i = { - /* .free_buffer = */ ggml_backend_cpu_buffer_free_buffer, - /* .get_base = */ ggml_backend_cpu_buffer_get_base, - /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes - /* .init_tensor = */ NULL, // no initialization required - /* .free_tensor = */ NULL, // no cleanup required -}; +void ggml_backend_event_wait(ggml_backend_t backend, ggml_backend_event_t event) { + GGML_ASSERT(backend->iface.event_wait != NULL); -// for buffers from ptr, free is not called -static struct ggml_backend_buffer_i cpu_backend_buffer_i_from_ptr = { - /* .free_buffer = */ NULL, // ptr is not owned by the buffer, so it does not need to be freed - /* .get_base = */ ggml_backend_cpu_buffer_get_base, - /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes - /* .init_tensor = */ NULL, - /* .free_tensor = */ NULL, + backend->iface.event_wait(backend, event); +} + +// backend registry + +#define GGML_REG_MAX_BACKENDS 16 + +struct ggml_backend_reg { + char name[128]; + ggml_backend_init_fn init_fn; + ggml_backend_buffer_type_t default_buffer_type; + void * user_data; }; -static const size_t TENSOR_ALIGNMENT = 64; // should be enough for AVX 512 +static struct ggml_backend_reg ggml_backend_registry[GGML_REG_MAX_BACKENDS]; +static size_t ggml_backend_registry_count = 0; -static ggml_backend_buffer_t ggml_backend_cpu_alloc_buffer(ggml_backend_t backend, size_t size) { - size += TENSOR_ALIGNMENT; // malloc may return an address that is not aligned - void * data = malloc(size); // TODO: maybe use GGML_ALIGNED_MALLOC? +GGML_CALL static ggml_backend_t ggml_backend_reg_cpu_init(const char * params, void * user_data); + +GGML_CALL static void ggml_backend_registry_init(void) { + static bool initialized = false; - GGML_ASSERT(data != NULL && "failed to allocate buffer"); + if (initialized) { + return; + } + + initialized = true; + + ggml_backend_register("CPU", ggml_backend_reg_cpu_init, ggml_backend_cpu_buffer_type(), NULL); + + // add forward decls here to avoid including the backend headers +#ifdef GGML_USE_CUDA + extern GGML_CALL void ggml_backend_cuda_reg_devices(void); + ggml_backend_cuda_reg_devices(); +#endif + +#ifdef GGML_USE_SYCL + extern void ggml_backend_sycl_reg_devices(void); + ggml_backend_sycl_reg_devices(); +#endif - return ggml_backend_buffer_init(backend, cpu_backend_buffer_i, data, size); +#ifdef GGML_USE_METAL + extern GGML_CALL ggml_backend_t ggml_backend_reg_metal_init(const char * params, void * user_data); + extern GGML_CALL ggml_backend_buffer_type_t ggml_backend_metal_buffer_type(void); + ggml_backend_register("Metal", ggml_backend_reg_metal_init, ggml_backend_metal_buffer_type(), NULL); +#endif + +#ifdef GGML_USE_VULKAN + extern GGML_CALL int ggml_backend_vk_reg_devices(void); + ggml_backend_vk_reg_devices(); +#endif + +#ifdef GGML_USE_KOMPUTE + extern GGML_CALL void ggml_backend_kompute_reg_devices(void); + ggml_backend_kompute_reg_devices(); +#endif } -static size_t ggml_backend_cpu_get_alignment(ggml_backend_t backend) { - return TENSOR_ALIGNMENT; - UNUSED(backend); +GGML_CALL void ggml_backend_register(const char * name, ggml_backend_init_fn init_fn, ggml_backend_buffer_type_t default_buffer_type, void * user_data) { + GGML_ASSERT(ggml_backend_registry_count < GGML_REG_MAX_BACKENDS); + + size_t id = ggml_backend_registry_count; + + ggml_backend_registry[id] = (struct ggml_backend_reg) { + /* .name = */ {0}, + /* .fn = */ init_fn, + /* .default_buffer_type = */ default_buffer_type, + /* .user_data = */ user_data, + }; + + snprintf(ggml_backend_registry[id].name, sizeof(ggml_backend_registry[id].name), "%s", name); + +#ifndef NDEBUG + fprintf(stderr, "%s: registered backend %s\n", __func__, name); +#endif + + ggml_backend_registry_count++; } -static void ggml_backend_cpu_set_tensor_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { - GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor write out of bounds"); - GGML_ASSERT(tensor->data != NULL && "tensor not allocated"); +size_t ggml_backend_reg_get_count(void) { + ggml_backend_registry_init(); - memcpy((char *)tensor->data + offset, data, size); + return ggml_backend_registry_count; +} + +size_t ggml_backend_reg_find_by_name(const char * name) { + ggml_backend_registry_init(); + + for (size_t i = 0; i < ggml_backend_registry_count; i++) { + // TODO: case insensitive in a portable way + if (strcmp(ggml_backend_registry[i].name, name) == 0) { + return i; + } + } - UNUSED(backend); + // not found + return SIZE_MAX; } -static void ggml_backend_cpu_get_tensor_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { - GGML_ASSERT(offset + size <= ggml_nbytes(tensor) && "tensor read out of bounds"); - GGML_ASSERT(tensor->data != NULL && "tensor not allocated"); +// init from backend:params string +ggml_backend_t ggml_backend_reg_init_backend_from_str(const char * backend_str) { + ggml_backend_registry_init(); - memcpy(data, (const char *)tensor->data + offset, size); + const char * params = strchr(backend_str, ':'); + char backend_name[128]; + if (params == NULL) { + snprintf(backend_name, sizeof(backend_name), "%s", backend_str); + params = ""; + } else { + snprintf(backend_name, sizeof(backend_name), "%.*s", (int)(params - backend_str), backend_str); + params++; + } + + size_t backend_i = ggml_backend_reg_find_by_name(backend_name); - UNUSED(backend); + if (backend_i == SIZE_MAX) { + fprintf(stderr, "%s: backend %s not found\n", __func__, backend_name); + return NULL; + } + + return ggml_backend_reg_init_backend(backend_i, params); } -static void ggml_backend_cpu_synchronize(ggml_backend_t backend) { - UNUSED(backend); +const char * ggml_backend_reg_get_name(size_t i) { + ggml_backend_registry_init(); + + GGML_ASSERT(i < ggml_backend_registry_count); + return ggml_backend_registry[i].name; } -static void ggml_backend_cpu_cpy_tensor_from(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst) { - ggml_backend_tensor_get(src, dst->data, 0, ggml_nbytes(src)); +ggml_backend_t ggml_backend_reg_init_backend(size_t i, const char * params) { + ggml_backend_registry_init(); - UNUSED(backend); + GGML_ASSERT(i < ggml_backend_registry_count); + return ggml_backend_registry[i].init_fn(params, ggml_backend_registry[i].user_data); } -static void ggml_backend_cpu_cpy_tensor_to(ggml_backend_t backend, struct ggml_tensor * src, struct ggml_tensor * dst) { - ggml_backend_tensor_set(dst, src->data, 0, ggml_nbytes(src)); +ggml_backend_buffer_type_t ggml_backend_reg_get_default_buffer_type(size_t i) { + ggml_backend_registry_init(); - UNUSED(backend); + GGML_ASSERT(i < ggml_backend_registry_count); + return ggml_backend_registry[i].default_buffer_type; } -struct ggml_backend_plan_cpu { - struct ggml_cplan cplan; - struct ggml_cgraph cgraph; -}; +ggml_backend_buffer_t ggml_backend_reg_alloc_buffer(size_t i, size_t size) { + ggml_backend_registry_init(); -static ggml_backend_graph_plan_t ggml_backend_cpu_graph_plan_create(ggml_backend_t backend, struct ggml_cgraph * cgraph) { - struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + GGML_ASSERT(i < ggml_backend_registry_count); + return ggml_backend_buft_alloc_buffer(ggml_backend_registry[i].default_buffer_type, size); +} - struct ggml_backend_plan_cpu * cpu_plan = malloc(sizeof(struct ggml_backend_plan_cpu)); +// backend CPU - cpu_plan->cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads); - cpu_plan->cgraph = *cgraph; +static const size_t TENSOR_ALIGNMENT = 32; // required for mmap as gguf only guarantees 32-byte alignment - if (cpu_plan->cplan.work_size > 0) { - cpu_plan->cplan.work_data = malloc(cpu_plan->cplan.work_size); - } +GGML_CALL static const char * ggml_backend_cpu_buffer_name(ggml_backend_buffer_t buffer) { + return "CPU"; - return cpu_plan; + GGML_UNUSED(buffer); } -static void ggml_backend_cpu_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { - struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; +GGML_CALL static void * ggml_backend_cpu_buffer_get_base(ggml_backend_buffer_t buffer) { + uintptr_t data = (uintptr_t)buffer->context; - free(cpu_plan->cplan.work_data); - free(cpu_plan); + // align the buffer + if (data % TENSOR_ALIGNMENT != 0) { + data = GGML_PAD(data, TENSOR_ALIGNMENT); + } - UNUSED(backend); + return (void *)data; } -static void ggml_backend_cpu_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { - struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; +GGML_CALL static void ggml_backend_cpu_buffer_free_buffer(ggml_backend_buffer_t buffer) { + free(buffer->context); +} - ggml_graph_compute(&cpu_plan->cgraph, &cpu_plan->cplan); +GGML_CALL static void ggml_backend_cpu_buffer_set_tensor(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size) { + memcpy((char *)tensor->data + offset, data, size); - UNUSED(backend); + GGML_UNUSED(buffer); } -static void ggml_backend_cpu_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { - struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; +GGML_CALL static void ggml_backend_cpu_buffer_get_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size) { + memcpy(data, (const char *)tensor->data + offset, size); - struct ggml_cplan cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads); + GGML_UNUSED(buffer); +} - if (cpu_ctx->work_size < cplan.work_size) { - // TODO: may be faster to free and use malloc to avoid the copy - cpu_ctx->work_data = realloc(cpu_ctx->work_data, cplan.work_size); - cpu_ctx->work_size = cplan.work_size; +GGML_CALL static bool ggml_backend_cpu_buffer_cpy_tensor(ggml_backend_buffer_t buffer, const struct ggml_tensor * src, struct ggml_tensor * dst) { + if (ggml_backend_buffer_is_host(src->buffer)) { + memcpy(dst->data, src->data, ggml_nbytes(src)); + return true; } + return false; - cplan.work_data = cpu_ctx->work_data; - - ggml_graph_compute(cgraph, &cplan); + GGML_UNUSED(buffer); } -static bool ggml_backend_cpu_supports_op(ggml_backend_t backend, const struct ggml_tensor * op) { - return true; - UNUSED(backend); - UNUSED(op); +GGML_CALL static void ggml_backend_cpu_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { + memset(buffer->context, value, buffer->size); } -static struct ggml_backend_i cpu_backend_i = { - /* .get_name = */ ggml_backend_cpu_name, - /* .free = */ ggml_backend_cpu_free, - /* .alloc_buffer = */ ggml_backend_cpu_alloc_buffer, - /* .get_alignment = */ ggml_backend_cpu_get_alignment, - /* .set_tensor_async = */ ggml_backend_cpu_set_tensor_async, - /* .get_tensor_async = */ ggml_backend_cpu_get_tensor_async, - /* .synchronize = */ ggml_backend_cpu_synchronize, - /* .cpy_tensor_from = */ ggml_backend_cpu_cpy_tensor_from, - /* .cpy_tensor_to = */ ggml_backend_cpu_cpy_tensor_to, - /* .graph_plan_create = */ ggml_backend_cpu_graph_plan_create, - /* .graph_plan_free = */ ggml_backend_cpu_graph_plan_free, - /* .graph_plan_compute = */ ggml_backend_cpu_graph_plan_compute, - /* .graph_compute = */ ggml_backend_cpu_graph_compute, - /* .supports_op = */ ggml_backend_cpu_supports_op, +static struct ggml_backend_buffer_i cpu_backend_buffer_i = { + /* .get_name = */ ggml_backend_cpu_buffer_name, + /* .free_buffer = */ ggml_backend_cpu_buffer_free_buffer, + /* .get_base = */ ggml_backend_cpu_buffer_get_base, + /* .init_tensor = */ NULL, // no initialization required + /* .set_tensor = */ ggml_backend_cpu_buffer_set_tensor, + /* .get_tensor = */ ggml_backend_cpu_buffer_get_tensor, + /* .cpy_tensor = */ ggml_backend_cpu_buffer_cpy_tensor, + /* .clear = */ ggml_backend_cpu_buffer_clear, + /* .reset = */ NULL, }; -ggml_backend_t ggml_backend_cpu_init(void) { - struct ggml_backend_cpu_context * ctx = malloc(sizeof(struct ggml_backend_cpu_context)); - - ctx->n_threads = GGML_DEFAULT_N_THREADS; - ctx->work_data = NULL; - ctx->work_size = 0; +// for buffers from ptr, free is not called +static struct ggml_backend_buffer_i cpu_backend_buffer_i_from_ptr = { + /* .get_name = */ ggml_backend_cpu_buffer_name, + /* .free_buffer = */ NULL, // ptr is not owned by the buffer, so it does not need to be freed + /* .get_base = */ ggml_backend_cpu_buffer_get_base, + /* .init_tensor = */ NULL, // no initialization required + /* .set_tensor = */ ggml_backend_cpu_buffer_set_tensor, + /* .get_tensor = */ ggml_backend_cpu_buffer_get_tensor, + /* .cpy_tensor = */ ggml_backend_cpu_buffer_cpy_tensor, + /* .clear = */ ggml_backend_cpu_buffer_clear, + /* .reset = */ NULL, +}; - ggml_backend_t cpu_backend = malloc(sizeof(struct ggml_backend)); +GGML_CALL static const char * ggml_backend_cpu_buffer_type_get_name(ggml_backend_buffer_type_t buft) { + return "CPU"; - *cpu_backend = (struct ggml_backend) { - /* .interface = */ cpu_backend_i, - /* .context = */ ctx - }; - return cpu_backend; + GGML_UNUSED(buft); } -bool ggml_backend_is_cpu(ggml_backend_t backend) { - return backend->iface.get_name == ggml_backend_cpu_name; +GGML_CALL static ggml_backend_buffer_t ggml_backend_cpu_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + size += TENSOR_ALIGNMENT; // malloc may return an address that is not aligned + void * data = malloc(size); // TODO: use GGML_ALIGNED_MALLOC (move to ggml-impl.h) + if (data == NULL) { + fprintf(stderr, "%s: failed to allocate buffer of size %zu\n", __func__, size); + return NULL; + } + + return ggml_backend_buffer_init(buft, cpu_backend_buffer_i, data, size); } -void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads) { - GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); +GGML_CALL static size_t ggml_backend_cpu_buffer_type_get_alignment(ggml_backend_buffer_type_t buft) { + return TENSOR_ALIGNMENT; - struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; - ctx->n_threads = n_threads; + GGML_UNUSED(buft); } -ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(ggml_backend_t backend_cpu, void * ptr, size_t size) { - return ggml_backend_buffer_init(backend_cpu, cpu_backend_buffer_i_from_ptr, ptr, size); +GGML_CALL static bool ggml_backend_cpu_buffer_type_supports_backend(ggml_backend_buffer_type_t buft, ggml_backend_t backend) { + return ggml_backend_is_cpu(backend); + + GGML_UNUSED(buft); } -// scheduler +GGML_CALL static bool ggml_backend_cpu_buffer_type_is_host(ggml_backend_buffer_type_t buft) { + return true; -#define GGML_MAX_BACKENDS 4 -#define GGML_MAX_SPLITS 256 -#define GGML_MAX_SPLIT_INPUTS 16 + GGML_UNUSED(buft); +} -struct ggml_backend_sched_split { - ggml_tallocr_t tallocr; - int i_start; - int i_end; - struct ggml_tensor * inputs[GGML_MAX_SPLIT_INPUTS]; - int n_inputs; - struct ggml_cgraph * graph; -}; +GGML_CALL ggml_backend_buffer_type_t ggml_backend_cpu_buffer_type(void) { + static struct ggml_backend_buffer_type ggml_backend_cpu_buffer_type = { + /* .iface = */ { + /* .get_name = */ ggml_backend_cpu_buffer_type_get_name, + /* .alloc_buffer = */ ggml_backend_cpu_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_cpu_buffer_type_get_alignment, + /* .get_max_size = */ NULL, // defaults to SIZE_MAX + /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes + /* .supports_backend = */ ggml_backend_cpu_buffer_type_supports_backend, + /* .is_host = */ ggml_backend_cpu_buffer_type_is_host, + }, + /* .context = */ NULL, + }; -struct ggml_backend_sched { - int n_backends; - ggml_backend_t backends[GGML_MAX_BACKENDS]; - ggml_tallocr_t tallocs[GGML_MAX_BACKENDS]; + return &ggml_backend_cpu_buffer_type; +} - ggml_gallocr_t galloc; +#ifdef GGML_USE_CPU_HBM - struct ggml_hash_set hash_set; - ggml_tallocr_t * node_talloc; // [hash_set.size] - struct ggml_tensor * (* node_copies)[GGML_MAX_BACKENDS]; // [hash_set.size][GGML_MAX_BACKENDS] +// buffer type HBM - struct ggml_cgraph * graph; - struct ggml_backend_sched_split splits[GGML_MAX_SPLITS]; - int n_splits; +#include - struct ggml_context * ctx; +GGML_CALL static const char * ggml_backend_cpu_hbm_buffer_type_get_name(ggml_backend_buffer_type_t buft) { + return "CPU_HBM"; - // align context_buffer to GGML_MEM_ALIGN - #ifdef _MSC_VER - __declspec(align(GGML_MEM_ALIGN)) - #else - __attribute__((aligned(GGML_MEM_ALIGN))) - #endif - char context_buffer[GGML_MAX_SPLITS*GGML_MAX_SPLIT_INPUTS*sizeof(struct ggml_tensor) + GGML_MAX_SPLITS*sizeof(struct ggml_cgraph)]; -}; + GGML_UNUSED(buft); +} -#define hash_id(node) ggml_hash_find_or_insert(sched->hash_set, node) -#define node_allocr(node) sched->node_talloc[hash_id(node)] +GGML_CALL static const char * ggml_backend_cpu_hbm_buffer_get_name(ggml_backend_buffer_t buf) { + return "CPU_HBM"; -static bool ggml_is_view_op(enum ggml_op op) { - return op == GGML_OP_VIEW || op == GGML_OP_RESHAPE || op == GGML_OP_PERMUTE || op == GGML_OP_TRANSPOSE; + GGML_UNUSED(buf); } -// returns the priority of the backend, lower is better -static int sched_backend_prio(ggml_backend_sched_t sched, ggml_backend_t backend) { - for (int i = 0; i < sched->n_backends; i++) { - if (sched->backends[i] == backend) { - return i; - } - } - return INT_MAX; +GGML_CALL static void ggml_backend_cpu_hbm_buffer_free_buffer(ggml_backend_buffer_t buffer) { + hbw_free(buffer->context); } -static int sched_allocr_prio(ggml_backend_sched_t sched, ggml_tallocr_t allocr) { - for (int i = 0; i < sched->n_backends; i++) { - if (sched->tallocs[i] == allocr) { - return i; - } +GGML_CALL static ggml_backend_buffer_t ggml_backend_cpu_hbm_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) { + //void * ptr = hbw_malloc(size); + void * ptr; + int result = hbw_posix_memalign(&ptr, ggml_backend_cpu_buffer_type_get_alignment(buft), size); + if (result != 0) { + fprintf(stderr, "failed to allocate HBM buffer of size %zu\n", size); + return NULL; } - return INT_MAX; -} + + ggml_backend_buffer_t buffer = ggml_backend_cpu_buffer_from_ptr(ptr, size); + buffer->buft = buft; + buffer->iface.get_name = ggml_backend_cpu_hbm_buffer_get_name; + buffer->iface.free_buffer = ggml_backend_cpu_hbm_buffer_free_buffer; + + return buffer; +} + +ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void) { + static struct ggml_backend_buffer_type ggml_backend_cpu_buffer_type_hbm = { + /* .iface = */ { + /* .get_name = */ ggml_backend_cpu_hbm_buffer_type_get_name, + /* .alloc_buffer = */ ggml_backend_cpu_hbm_buffer_type_alloc_buffer, + /* .get_alignment = */ ggml_backend_cpu_buffer_type_get_alignment, + /* .get_max_size = */ NULL, // defaults to SIZE_MAX + /* .get_alloc_size = */ NULL, // defaults to ggml_nbytes + /* .supports_backend = */ ggml_backend_cpu_buffer_type_supports_backend, + /* .is_host = */ ggml_backend_cpu_buffer_type_is_host, + }, + /* .context = */ NULL, + }; + + return &ggml_backend_cpu_buffer_type_hbm; +} +#endif + +struct ggml_backend_cpu_context { + int n_threads; + void * work_data; + size_t work_size; + + ggml_abort_callback abort_callback; + void * abort_callback_data; +}; + +GGML_CALL static const char * ggml_backend_cpu_name(ggml_backend_t backend) { + return "CPU"; + + GGML_UNUSED(backend); +} + +GGML_CALL static void ggml_backend_cpu_free(ggml_backend_t backend) { + struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + free(cpu_ctx->work_data); + free(cpu_ctx); + free(backend); +} + +GGML_CALL static ggml_backend_buffer_type_t ggml_backend_cpu_get_default_buffer_type(ggml_backend_t backend) { + return ggml_backend_cpu_buffer_type(); + + GGML_UNUSED(backend); +} + +struct ggml_backend_plan_cpu { + struct ggml_cplan cplan; + struct ggml_cgraph cgraph; +}; + +GGML_CALL static ggml_backend_graph_plan_t ggml_backend_cpu_graph_plan_create(ggml_backend_t backend, const struct ggml_cgraph * cgraph) { + struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + + struct ggml_backend_plan_cpu * cpu_plan = malloc(sizeof(struct ggml_backend_plan_cpu)); + + cpu_plan->cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads); + cpu_plan->cgraph = *cgraph; // FIXME: deep copy + + if (cpu_plan->cplan.work_size > 0) { + cpu_plan->cplan.work_data = malloc(cpu_plan->cplan.work_size); + if (cpu_plan->cplan.work_data == NULL) { + free(cpu_plan); + return NULL; + } + } + + cpu_plan->cplan.abort_callback = cpu_ctx->abort_callback; + cpu_plan->cplan.abort_callback_data = cpu_ctx->abort_callback_data; + + return cpu_plan; +} + +GGML_CALL static void ggml_backend_cpu_graph_plan_free(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { + struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; + + free(cpu_plan->cplan.work_data); + free(cpu_plan); + + GGML_UNUSED(backend); +} + +GGML_CALL static enum ggml_status ggml_backend_cpu_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan) { + struct ggml_backend_plan_cpu * cpu_plan = (struct ggml_backend_plan_cpu *)plan; + + return ggml_graph_compute(&cpu_plan->cgraph, &cpu_plan->cplan); + + GGML_UNUSED(backend); +} + +GGML_CALL static enum ggml_status ggml_backend_cpu_graph_compute(ggml_backend_t backend, struct ggml_cgraph * cgraph) { + struct ggml_backend_cpu_context * cpu_ctx = (struct ggml_backend_cpu_context *)backend->context; + + struct ggml_cplan cplan = ggml_graph_plan(cgraph, cpu_ctx->n_threads); + + if (cpu_ctx->work_size < cplan.work_size) { + free(cpu_ctx->work_data); + cpu_ctx->work_data = malloc(cplan.work_size); + if (cpu_ctx->work_data == NULL) { + cpu_ctx->work_size = 0; + return GGML_STATUS_ALLOC_FAILED; + } + cpu_ctx->work_size = cplan.work_size; + } + cplan.work_data = cpu_ctx->work_data; + + cplan.abort_callback = cpu_ctx->abort_callback; + cplan.abort_callback_data = cpu_ctx->abort_callback_data; + + return ggml_graph_compute(cgraph, &cplan); +} + +GGML_CALL static bool ggml_backend_cpu_supports_op(ggml_backend_t backend, const struct ggml_tensor * op) { + switch (op->op) { + case GGML_OP_CPY: + return op->type != GGML_TYPE_IQ2_XXS && op->type != GGML_TYPE_IQ2_XS && op->type != GGML_TYPE_IQ1_S; // missing type_traits.from_float + case GGML_OP_MUL_MAT: + return op->src[1]->type == GGML_TYPE_F32 || op->src[1]->type == ggml_internal_get_type_traits(op->src[0]->type).vec_dot_type; + default: + return true; + } + + GGML_UNUSED(backend); +} + +static struct ggml_backend_i cpu_backend_i = { + /* .get_name = */ ggml_backend_cpu_name, + /* .free = */ ggml_backend_cpu_free, + /* .get_default_buffer_type = */ ggml_backend_cpu_get_default_buffer_type, + /* .set_tensor_async = */ NULL, + /* .get_tensor_async = */ NULL, + /* .cpy_tensor_async = */ NULL, + /* .synchronize = */ NULL, + /* .graph_plan_create = */ ggml_backend_cpu_graph_plan_create, + /* .graph_plan_free = */ ggml_backend_cpu_graph_plan_free, + /* .graph_plan_compute = */ ggml_backend_cpu_graph_plan_compute, + /* .graph_compute = */ ggml_backend_cpu_graph_compute, + /* .supports_op = */ ggml_backend_cpu_supports_op, + /* .offload_op = */ NULL, + /* .event_new = */ NULL, + /* .event_free = */ NULL, + /* .event_record = */ NULL, + /* .event_wait = */ NULL, + /* .event_synchronize = */ NULL, +}; + +static ggml_guid_t ggml_backend_cpu_guid(void) { + static ggml_guid guid = { 0xaa, 0x67, 0xc7, 0x43, 0x96, 0xe6, 0xa3, 0x8a, 0xe3, 0xaf, 0xea, 0x92, 0x36, 0xbc, 0xfc, 0x89 }; + return &guid; +} + +ggml_backend_t ggml_backend_cpu_init(void) { + struct ggml_backend_cpu_context * ctx = malloc(sizeof(struct ggml_backend_cpu_context)); + if (ctx == NULL) { + return NULL; + } + + ctx->n_threads = GGML_DEFAULT_N_THREADS; + ctx->work_data = NULL; + ctx->work_size = 0; + ctx->abort_callback = NULL; + ctx->abort_callback_data = NULL; + + ggml_backend_t cpu_backend = malloc(sizeof(struct ggml_backend)); + if (cpu_backend == NULL) { + free(ctx); + return NULL; + } + + *cpu_backend = (struct ggml_backend) { + /* .guid = */ ggml_backend_cpu_guid(), + /* .interface = */ cpu_backend_i, + /* .context = */ ctx + }; + return cpu_backend; +} + +GGML_CALL bool ggml_backend_is_cpu(ggml_backend_t backend) { + return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_cpu_guid()); +} + +void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads) { + GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); + + struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; + ctx->n_threads = n_threads; +} + +void ggml_backend_cpu_set_abort_callback(ggml_backend_t backend_cpu, ggml_abort_callback abort_callback, void * abort_callback_data) { + GGML_ASSERT(ggml_backend_is_cpu(backend_cpu)); + + struct ggml_backend_cpu_context * ctx = (struct ggml_backend_cpu_context *)backend_cpu->context; + ctx->abort_callback = abort_callback; + ctx->abort_callback_data = abort_callback_data; +} + +GGML_CALL ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(void * ptr, size_t size) { + GGML_ASSERT((uintptr_t)ptr % TENSOR_ALIGNMENT == 0 && "buffer pointer must be aligned"); + return ggml_backend_buffer_init(ggml_backend_cpu_buffer_type(), cpu_backend_buffer_i_from_ptr, ptr, size); +} + +GGML_CALL static ggml_backend_t ggml_backend_reg_cpu_init(const char * params, void * user_data) { + return ggml_backend_cpu_init(); + + GGML_UNUSED(params); + GGML_UNUSED(user_data); +} + +// multi-buffer buffer + +struct ggml_backend_multi_buffer_context { + ggml_backend_buffer_t * buffers; + size_t n_buffers; +}; + +typedef struct ggml_backend_multi_buffer_context * ggml_backend_multi_buffer_context_t; + +GGML_CALL static const char * ggml_backend_multi_buffer_get_name(ggml_backend_buffer_t buffer) { + ggml_backend_multi_buffer_context_t ctx = (ggml_backend_multi_buffer_context_t) buffer->context; + + return ctx->buffers[0]->iface.get_name(ctx->buffers[0]); +} + +GGML_CALL static void ggml_backend_multi_buffer_free_buffer(ggml_backend_buffer_t buffer) { + ggml_backend_multi_buffer_context_t ctx = (ggml_backend_multi_buffer_context_t) buffer->context; + for (size_t i = 0; i < ctx->n_buffers; i++) { + ggml_backend_buffer_free(ctx->buffers[i]); + } + + free(ctx->buffers); + free(ctx); +} + +GGML_CALL static void ggml_backend_multi_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) { + ggml_backend_multi_buffer_context_t ctx = (ggml_backend_multi_buffer_context_t) buffer->context; + for (size_t i = 0; i < ctx->n_buffers; i++) { + ggml_backend_buffer_clear(ctx->buffers[i], value); + } +} + +static struct ggml_backend_buffer_i ggml_backend_multi_buffer_context_interface(void) { + static struct ggml_backend_buffer_i multi_backend_buffer_i = { + /* .get_name = */ ggml_backend_multi_buffer_get_name, + /* .free_buffer = */ ggml_backend_multi_buffer_free_buffer, + /* .get_base = */ NULL, + /* .init_tensor = */ NULL, + /* .set_tensor = */ NULL, + /* .get_tensor = */ NULL, + /* .cpy_tensor = */ NULL, + /* .clear = */ ggml_backend_multi_buffer_clear, + /* .reset = */ NULL, + }; + + return multi_backend_buffer_i; +} + +GGML_CALL ggml_backend_buffer_t ggml_backend_multi_buffer_alloc_buffer(ggml_backend_buffer_t * buffers, size_t n_buffers) { + ggml_backend_multi_buffer_context_t ctx = (ggml_backend_multi_buffer_context_t) malloc(sizeof(struct ggml_backend_multi_buffer_context)); + ctx->n_buffers = n_buffers; + ctx->buffers = (ggml_backend_buffer_t *) malloc(n_buffers * sizeof(ggml_backend_buffer_t)); + + GGML_ASSERT(ctx->buffers != NULL); + + size_t total_size = 0; + for (size_t i = 0; i < n_buffers; i++) { + ctx->buffers[i] = buffers[i]; + total_size += ggml_backend_buffer_get_size(buffers[i]); + } + + return ggml_backend_buffer_init(buffers[0]->buft, ggml_backend_multi_buffer_context_interface(), ctx, total_size); +} + +GGML_CALL bool ggml_backend_buffer_is_multi_buffer(ggml_backend_buffer_t buffer) { + return buffer->iface.get_name == ggml_backend_multi_buffer_get_name; +} + +GGML_CALL void ggml_backend_multi_buffer_set_usage(ggml_backend_buffer_t buffer, enum ggml_backend_buffer_usage usage) { + GGML_ASSERT(ggml_backend_buffer_is_multi_buffer(buffer)); + ggml_backend_multi_buffer_context_t ctx = (ggml_backend_multi_buffer_context_t) buffer->context; + for (size_t i = 0; i < ctx->n_buffers; i++) { + ggml_backend_buffer_set_usage(ctx->buffers[i], usage); + } +} + +// creates a copy of the tensor with the same memory layout +static struct ggml_tensor * ggml_dup_tensor_layout(struct ggml_context * ctx, const struct ggml_tensor * tensor) { + struct ggml_tensor * dup = ggml_dup_tensor(ctx, tensor); + for (int i = 0; i < GGML_MAX_DIMS; i++) { + dup->nb[i] = tensor->nb[i]; + } + return dup; +} + +static bool ggml_is_view_op(enum ggml_op op) { + return op == GGML_OP_VIEW || op == GGML_OP_RESHAPE || op == GGML_OP_PERMUTE || op == GGML_OP_TRANSPOSE; +} + +// scheduler + +#ifndef GGML_SCHED_MAX_BACKENDS +#define GGML_SCHED_MAX_BACKENDS 16 +#endif + +#ifndef GGML_SCHED_MAX_SPLITS +#define GGML_SCHED_MAX_SPLITS 2048 +#endif + +#ifndef GGML_SCHED_MAX_SPLIT_INPUTS +#define GGML_SCHED_MAX_SPLIT_INPUTS GGML_MAX_SRC +#endif + +#ifndef GGML_SCHED_MAX_COPIES +#define GGML_SCHED_MAX_COPIES 4 +#endif + +struct ggml_backend_sched_split { + int backend_id; + int i_start; + int i_end; + struct ggml_tensor * inputs[GGML_SCHED_MAX_SPLIT_INPUTS]; + int n_inputs; + // graph view of this split + struct ggml_cgraph graph; +}; + +struct ggml_backend_sched { + bool is_reset; // true if the scheduler has been reset since the last graph split + bool is_alloc; + + int n_backends; + + ggml_backend_t backends[GGML_SCHED_MAX_BACKENDS]; + ggml_backend_buffer_type_t bufts[GGML_SCHED_MAX_BACKENDS]; + ggml_gallocr_t galloc; + + // hash keys of the nodes in the graph + struct ggml_hash_set hash_set; + // hash values + int * tensor_backend_id; + struct ggml_tensor * (* tensor_copies)[GGML_SCHED_MAX_BACKENDS][GGML_SCHED_MAX_COPIES]; + + int * node_backend_ids; // [graph_size] + int * leaf_backend_ids; // [graph_size] + + // copy of the graph with modified inputs + struct ggml_cgraph * graph; + + // graph splits + struct ggml_backend_sched_split * splits; + int n_splits; + int splits_capacity; + + // pipeline parallelism support + int n_copies; + int cur_copy; + ggml_backend_event_t events[GGML_SCHED_MAX_BACKENDS][GGML_SCHED_MAX_COPIES]; + struct ggml_tensor * graph_inputs[GGML_SCHED_MAX_SPLIT_INPUTS]; + int n_graph_inputs; + + struct ggml_context * ctx; + + ggml_backend_sched_eval_callback callback_eval; + void * callback_eval_user_data; + + // align context_buffer to GGML_MEM_ALIGN +#ifdef _MSC_VER + __declspec(align(GGML_MEM_ALIGN)) +#else + __attribute__((aligned(GGML_MEM_ALIGN))) +#endif + char context_buffer[GGML_SCHED_MAX_SPLITS*GGML_SCHED_MAX_SPLIT_INPUTS*2*sizeof(struct ggml_tensor) + sizeof(struct ggml_cgraph)]; +}; + +#define hash_id(tensor) ggml_hash_find_or_insert(sched->hash_set, tensor) +#define tensor_backend_id(tensor) sched->tensor_backend_id[hash_id(tensor)] + +// returns the priority of the backend, lower id is higher priority +static int ggml_backend_sched_backend_id(ggml_backend_sched_t sched, ggml_backend_t backend) { + for (int i = 0; i < sched->n_backends; i++) { + if (sched->backends[i] == backend) { + return i; + } + } + return -1; +} + +static int ggml_backend_sched_backend_from_buffer(ggml_backend_sched_t sched, const struct ggml_tensor * tensor) { + ggml_backend_buffer_t buffer = tensor->buffer; + if (buffer == NULL) { + return -1; + } + + // find highest prio backend that supports the buffer type + for (int i = 0; i < sched->n_backends; i++) { + if (ggml_backend_buft_supports_backend(buffer->buft, sched->backends[i])) { + return i; + } + } + + fprintf(stderr, "%s: error: no backend supports buffer type %s used in tensor %s\n", + __func__, ggml_backend_buffer_name(buffer), tensor->name); + GGML_ASSERT(false); + + return -1; +} + +#if 0 +static char causes[GGML_DEFAULT_GRAPH_SIZE*16 + GGML_SCHED_MAX_SPLITS*GGML_SCHED_MAX_SPLIT_INPUTS][128]; // debug only +#define SET_CAUSE(node, ...) sprintf(causes[hash_id(node)], __VA_ARGS__) +#define GET_CAUSE(node) causes[hash_id(node)] +#else +#define SET_CAUSE(node, ...) +#define GET_CAUSE(node) "" +#endif // returns the backend that should be used for the node based on the current locations -char causes[GGML_DEFAULT_GRAPH_SIZE*4 + GGML_MAX_SPLITS*GGML_MAX_SPLIT_INPUTS][128]; // debug, remove -static ggml_backend_t sched_backend_from_cur(ggml_backend_sched_t sched, struct ggml_tensor * node) { - // if the dst tensor is already allocated in a buffer, we must assume that it is critical to keep it there - // ie. kv cache updates - // note that this doesn't allow fallback to CPU. need to add output tensors to the splits to copy the data back to the original backend. - // dst - ggml_backend_t cur_backend = ggml_get_backend(node); - if (cur_backend != NULL) { - sprintf(causes[hash_id(node)], "1.dst"); - return cur_backend; +static int ggml_backend_sched_backend_id_from_cur(ggml_backend_sched_t sched, struct ggml_tensor * tensor) { + // TODO: use supports_op to check if the backend supports the op + + // assign pre-allocated nodes to their backend + int cur_backend_id = ggml_backend_sched_backend_from_buffer(sched, tensor); + if (cur_backend_id != -1) { + SET_CAUSE(tensor, "1.dst"); + return cur_backend_id; } // view_src - if (node->view_src != NULL && ggml_get_backend(node->view_src) != NULL) { - sprintf(causes[hash_id(node)], "1.vsrc"); - return ggml_get_backend(node->view_src); + if (tensor->view_src != NULL) { + cur_backend_id = ggml_backend_sched_backend_from_buffer(sched, tensor->view_src); + if (cur_backend_id != -1) { + SET_CAUSE(tensor, "1.vsrc"); + return cur_backend_id; + } } - // src - int cur_prio = INT_MAX; - size_t cur_size = 0; + // graph input + if (tensor->flags & GGML_TENSOR_FLAG_INPUT) { + cur_backend_id = sched->n_backends - 1; // last backend (assumed CPU) + SET_CAUSE(tensor, "1.inp"); + return cur_backend_id; + } + // assign nodes that use weights to the backend of the weights + // operations with weights are preferably run on the same backend as the weights for (int i = 0; i < GGML_MAX_SRC; i++) { - const struct ggml_tensor * src = node->src[i]; + const struct ggml_tensor * src = tensor->src[i]; if (src == NULL) { - break; + continue; } - ggml_backend_t src_backend = ggml_get_backend(src); - if (src_backend != NULL) { - int src_prio = sched_backend_prio(sched, src_backend); - size_t src_size = ggml_nbytes(src); - if (src_prio < cur_prio && src_size >= cur_size) { - cur_prio = src_prio; - cur_size = src_size; - cur_backend = src_backend; - sprintf(causes[hash_id(node)], "1.src%d", i); + if (src->buffer != NULL && src->buffer->usage == GGML_BACKEND_BUFFER_USAGE_WEIGHTS) { + int src_backend_id = ggml_backend_sched_backend_from_buffer(sched, src); + // check if a backend with higher prio wants to offload the op + if (src_backend_id == sched->n_backends - 1) { + for (int b = 0; b < src_backend_id; b++) { + if (ggml_backend_offload_op(sched->backends[b], tensor)) { + SET_CAUSE(tensor, "1.off"); + return b; + } + } } + SET_CAUSE(tensor, "1.wgt%d", i); + return src_backend_id; } } - return cur_backend; + + return -1; } static char * fmt_size(size_t size) { @@ -535,14 +1185,16 @@ static char * fmt_size(size_t size) { return buffer; } -static void sched_print_assignments(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { +static void ggml_backend_sched_print_assignments(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { int cur_split = 0; for (int i = 0; i < graph->n_nodes; i++) { if (cur_split < sched->n_splits && i == sched->splits[cur_split].i_start) { - ggml_backend_t split_backend = ggml_tallocr_get_buffer(sched->splits[cur_split].tallocr)->backend; - fprintf(stderr, "\n## SPLIT #%d: %s # %d inputs: ", cur_split, ggml_backend_name(split_backend), sched->splits[cur_split].n_inputs); + ggml_backend_t split_backend = sched->backends[sched->splits[cur_split].backend_id]; + fprintf(stderr, "\n## SPLIT #%d: %s # %d inputs: ", cur_split, ggml_backend_name(split_backend), + sched->splits[cur_split].n_inputs); for (int j = 0; j < sched->splits[cur_split].n_inputs; j++) { - fprintf(stderr, "[%s (%5.5s)] ", sched->splits[cur_split].inputs[j]->name, fmt_size(ggml_nbytes(sched->splits[cur_split].inputs[j]))); + fprintf(stderr, "[%s (%5.5s)] ", sched->splits[cur_split].inputs[j]->name, + fmt_size(ggml_nbytes(sched->splits[cur_split].inputs[j]))); } fprintf(stderr, "\n"); cur_split++; @@ -551,341 +1203,558 @@ static void sched_print_assignments(ggml_backend_sched_t sched, struct ggml_cgra if (ggml_is_view_op(node->op)) { continue; } - ggml_tallocr_t node_allocr = node_allocr(node); - ggml_backend_t node_backend = node_allocr ? ggml_tallocr_get_buffer(node_allocr)->backend : NULL; - fprintf(stderr, "node #%3d (%10.10s): %20.20s (%4.4s) [%4.4s %8.8s]:", i, ggml_op_name(node->op), node->name, fmt_size(ggml_nbytes(node)), node_allocr ? ggml_backend_name(node_backend) : "NULL", causes[hash_id(node)]); + ggml_backend_t tensor_backend = ggml_backend_sched_get_tensor_backend(sched, node); + fprintf(stderr, "node #%3d (%10.10s): %20.20s (%5.5s) [%5.5s %8.8s]:", i, ggml_op_name(node->op), node->name, + fmt_size(ggml_nbytes(node)), tensor_backend ? ggml_backend_name(tensor_backend) : "NULL", GET_CAUSE(node)); for (int j = 0; j < GGML_MAX_SRC; j++) { struct ggml_tensor * src = node->src[j]; if (src == NULL) { - break; + continue; } - ggml_tallocr_t src_allocr = node_allocr(src); - ggml_backend_t src_backend = src_allocr ? ggml_tallocr_get_buffer(src_allocr)->backend : NULL; - fprintf(stderr, " %20.20s (%4.4s) [%4.4s %8.8s]", src->name, fmt_size(ggml_nbytes(src)), src_backend ? ggml_backend_name(src_backend) : "NULL", causes[hash_id(src)]); + ggml_backend_t src_backend = ggml_backend_sched_get_tensor_backend(sched, src); + fprintf(stderr, " %20.20s (%5.5s) [%5.5s %8.8s]", src->name, + fmt_size(ggml_nbytes(src)), src_backend ? ggml_backend_name(src_backend) : "NULL", GET_CAUSE(src)); } fprintf(stderr, "\n"); } } -// creates a copy of the tensor with the same memory layout -static struct ggml_tensor * ggml_dup_tensor_layout(struct ggml_context * ctx, const struct ggml_tensor * tensor) { - struct ggml_tensor * dup = ggml_dup_tensor(ctx, tensor); - for (int i = 0; i < GGML_MAX_DIMS; i++) { - dup->nb[i] = tensor->nb[i]; - } - return dup; -} +//#define DEBUG_PASS1 +//#define DEBUG_PASS2 +//#define DEBUG_PASS3 +//#define DEBUG_PASS4 // assigns backends to ops and splits the graph into subgraphs that can be computed on the same backend -// TODO: merge passes -static void sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { - // reset state - size_t hash_size = sched->hash_set.size; - memset(sched->hash_set.keys, 0, sizeof(sched->hash_set.keys[0]) * hash_size); - memset(sched->node_talloc, 0, sizeof(sched->node_talloc[0]) * hash_size); - memset(sched->node_copies, 0, sizeof(sched->node_copies[0]) * hash_size); +static void ggml_backend_sched_split_graph(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { + // reset splits sched->n_splits = 0; + sched->n_graph_inputs = 0; + sched->is_reset = false; struct ggml_init_params params = { - /*.mem_size = */ sizeof(sched->context_buffer), - /*.mem_buffer = */ sched->context_buffer, - /*.no_alloc = */ true + /* .mem_size = */ sizeof(sched->context_buffer), + /* .mem_buffer = */ sched->context_buffer, + /* .no_alloc = */ true }; - if (sched->ctx != NULL) { - ggml_free(sched->ctx); - } + ggml_free(sched->ctx); sched->ctx = ggml_init(params); + if (sched->ctx == NULL) { + fprintf(stderr, "%s: failed to initialize context\n", __func__); + GGML_ASSERT(false); + } - // pass 1: assign backends to ops with allocated inputs + // pass 1: assign backends to ops with pre-allocated inputs for (int i = 0; i < graph->n_leafs; i++) { struct ggml_tensor * leaf = graph->leafs[i]; - if (node_allocr(leaf) != NULL) { + int * leaf_backend_id = &tensor_backend_id(leaf); + if (*leaf_backend_id != -1) { // do not overwrite user assignments continue; } - ggml_backend_t leaf_backend = ggml_get_backend(leaf); - if (leaf_backend == NULL && leaf->view_src != NULL) { - leaf_backend = ggml_get_backend(leaf->view_src); - } - if (leaf_backend != NULL) { - node_allocr(leaf) = ggml_backend_sched_get_tallocr(sched, leaf_backend); - } + *leaf_backend_id = ggml_backend_sched_backend_id_from_cur(sched, leaf); } for (int i = 0; i < graph->n_nodes; i++) { struct ggml_tensor * node = graph->nodes[i]; - if (node_allocr(node) != NULL) { + int * node_backend_id = &tensor_backend_id(node); + if (*node_backend_id != -1) { // do not overwrite user assignments continue; } - ggml_backend_t node_backend = sched_backend_from_cur(sched, node); - if (node_backend != NULL) { - node_allocr(node) = ggml_backend_sched_get_tallocr(sched, node_backend); + *node_backend_id = ggml_backend_sched_backend_id_from_cur(sched, node); + // src + for (int j = 0; j < GGML_MAX_SRC; j++) { + struct ggml_tensor * src = node->src[j]; + if (src == NULL) { + continue; + } + int * src_backend_id = &tensor_backend_id(src); + if (*src_backend_id == -1) { + *src_backend_id = ggml_backend_sched_backend_id_from_cur(sched, src); + } } } - //printf("PASS 1 ASSIGNMENTS\n"); sched_print_assignments(sched, graph); +#ifdef DEBUG_PASS1 + fprintf(stderr, "PASS 1 ASSIGNMENTS\n"); ggml_backend_sched_print_assignments(sched, graph); +#endif - // pass 2: assign backends to ops from current assignments - // TODO: - // - reuse sched_backend_from_cur - for (int i = 0; i < graph->n_nodes; i++) { - struct ggml_tensor * node = graph->nodes[i]; - ggml_tallocr_t node_allocr = node_allocr(node); - if (node_allocr == NULL) { - int cur_prio = INT_MAX; - size_t cur_size = 0; - for (int j = 0; j < GGML_MAX_SRC; j++) { - struct ggml_tensor * src = node->src[j]; - if (src == NULL) { - break; + // pass 2: expand current backend assignments + // assign the same backend to adjacent nodes + // expand gpu backends (i.e. non last prio) up and down, ignoring cpu (the lowest priority backend) + // thus, cpu will never be used unless weights are on cpu, or there are no gpu ops between cpu ops + + + // pass 2.2 expand gpu down + { + int cur_backend_id = -1; + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + if (ggml_is_view_op(node->op)) { + continue; + } + int * node_backend_id = &tensor_backend_id(node); + if (*node_backend_id != -1) { + if (*node_backend_id == sched->n_backends - 1) { + // skip cpu (lowest prio backend) + cur_backend_id = -1; + } else { + cur_backend_id = *node_backend_id; } - ggml_tallocr_t src_allocr = node_allocr(src); - if (src_allocr != NULL) { - int src_prio = sched_allocr_prio(sched, src_allocr); - size_t src_size = ggml_nbytes(src); - if (src_prio < cur_prio && src_size >= cur_size) { - cur_prio = src_prio; - cur_size = src_size; - node_allocr = src_allocr; - sprintf(causes[hash_id(node)], "2.src%d", j); - } + } else { + *node_backend_id = cur_backend_id; + SET_CAUSE(node, "2.2"); + } + } + } + // pass 2.1 expand gpu up + { + int cur_backend_id = -1; + for (int i = graph->n_nodes - 1; i >= 0; i--) { + struct ggml_tensor * node = graph->nodes[i]; + if (ggml_is_view_op(node->op)) { + continue; + } + int * node_backend_id = &tensor_backend_id(node); + if (*node_backend_id != -1) { + if (*node_backend_id == sched->n_backends - 1) { + // skip cpu (lowest prio backend) + cur_backend_id = -1; + } else { + cur_backend_id = *node_backend_id; } + } else { + *node_backend_id = cur_backend_id; + SET_CAUSE(node, "2.1"); } - if (node_allocr != NULL) { - node_allocr(node) = node_allocr; + } + } + // pass 2.4 expand rest down + { + int cur_backend_id = -1; + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + if (ggml_is_view_op(node->op)) { + continue; + } + int * node_backend_id = &tensor_backend_id(node); + if (*node_backend_id != -1) { + cur_backend_id = *node_backend_id; + } else { + *node_backend_id = cur_backend_id; + SET_CAUSE(node, "2.4"); + } + } + } + // pass 2.3 expand rest up + { + int cur_backend_id = -1; + for (int i = graph->n_nodes - 1; i >= 0; i--) { + struct ggml_tensor * node = graph->nodes[i]; + if (ggml_is_view_op(node->op)) { + continue; + } + int * node_backend_id = &tensor_backend_id(node); + if (*node_backend_id != -1) { + cur_backend_id = *node_backend_id; + } else { + *node_backend_id = cur_backend_id; + SET_CAUSE(node, "2.3"); } } } - //printf("PASS 2 ASSIGNMENTS\n"); sched_print_assignments(sched, graph); - // pass 3: assign backends to remaining src from dst (should only be leafs) +#ifdef DEBUG_PASS2 + fprintf(stderr, "PASS 2 ASSIGNMENTS\n"); ggml_backend_sched_print_assignments(sched, graph); +#endif + + // pass 3: assign backends to remaining src from dst and view_src for (int i = 0; i < graph->n_nodes; i++) { struct ggml_tensor * node = graph->nodes[i]; - ggml_tallocr_t node_allocr = node_allocr(node); + int * cur_backend_id = &tensor_backend_id(node); + if (node->view_src != NULL && *cur_backend_id == -1) { + *cur_backend_id = tensor_backend_id(node->view_src); + SET_CAUSE(node, "3.vsrc"); + } for (int j = 0; j < GGML_MAX_SRC; j++) { struct ggml_tensor * src = node->src[j]; if (src == NULL) { - break; + continue; } - ggml_tallocr_t src_allocr = node_allocr(src); - if (src_allocr == NULL) { - node_allocr(src) = node_allocr; + int * src_backend_id = &tensor_backend_id(src); + if (*src_backend_id == -1) { + if (src->view_src != NULL) { + // views are always on the same backend as the source + *src_backend_id = tensor_backend_id(src->view_src); + SET_CAUSE(src, "3.vsrc"); + } else { + *src_backend_id = *cur_backend_id; + SET_CAUSE(src, "3.cur"); + } } } } - //printf("PASS 3 ASSIGNMENTS\n"); sched_print_assignments(sched, graph); +#ifdef DEBUG_PASS3 + fprintf(stderr, "PASS 3 ASSIGNMENTS\n"); ggml_backend_sched_print_assignments(sched, graph); +#endif // pass 4: split graph, find tensors that need to be copied - // TODO: - // - when switching from a less preferred backend to a more preferred backend, check if it is possible to move the switch to an earlier point for the same cost - // find first backend - int cur_split = 0; - for (int i = 0; i < graph->n_nodes; i++) { - struct ggml_tensor * node = graph->nodes[i]; - if (node->view_src == NULL) { - sched->splits[0].tallocr = node_allocr(node); - break; - } - } - sched->splits[0].i_start = 0; - sched->splits[0].n_inputs = 0; - memset(sched->splits[0].inputs, 0, sizeof(sched->splits[0].inputs)); //HACK - ggml_tallocr_t cur_allocr = sched->splits[0].tallocr; - size_t cur_backend_id = sched_allocr_prio(sched, cur_allocr); - for (int i = 0; i < graph->n_nodes; i++) { - struct ggml_tensor * node = graph->nodes[i]; - - if (ggml_is_view_op(node->op)) { - continue; + { + int i_split = 0; + struct ggml_backend_sched_split * split = &sched->splits[0]; + // find the backend of the first split, skipping view ops + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + if (!ggml_is_view_op(node->op)) { + split->backend_id = tensor_backend_id(node); + break; + } } + split->i_start = 0; + split->n_inputs = 0; + memset(split->inputs, 0, sizeof(split->inputs)); //HACK + int cur_backend_id = split->backend_id; + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + + if (ggml_is_view_op(node->op)) { + continue; + } - ggml_tallocr_t node_allocr = node_allocr(node); + const int node_backend_id = tensor_backend_id(node); - if (node_allocr != cur_allocr) { - sched->splits[cur_split].i_end = i; - cur_split++; - GGML_ASSERT(cur_split < GGML_MAX_SPLITS); - sched->splits[cur_split].tallocr = node_allocr; - sched->splits[cur_split].i_start = i; - sched->splits[cur_split].n_inputs = 0; - memset(sched->splits[cur_split].inputs, 0, sizeof(sched->splits[cur_split].inputs)); //HACK - cur_allocr = node_allocr; - cur_backend_id = sched_allocr_prio(sched, cur_allocr); - } + GGML_ASSERT(node_backend_id != -1); // all nodes should be assigned by now - // find inputs that are not on the same backend - for (int j = 0; j < GGML_MAX_SRC; j++) { - struct ggml_tensor * src = node->src[j]; - if (src == NULL) { - break; + // check if we should start a new split based on the sources of the current node + bool need_new_split = false; + if (node_backend_id == cur_backend_id && split->n_inputs > 0) { + for (int j = 0; j < GGML_MAX_SRC; j++) { + struct ggml_tensor * src = node->src[j]; + if (src == NULL) { + continue; + } + // check if a weight is on a different backend + // by starting a new split, the memory of the previously offloaded weights can be reused + if (src->buffer != NULL && src->buffer->usage == GGML_BACKEND_BUFFER_USAGE_WEIGHTS) { + int src_backend_id = tensor_backend_id(src); + if (src_backend_id != -1 && src_backend_id != cur_backend_id) { + need_new_split = true; + break; + } + } + // check if the split has too many inputs + if (split->n_inputs == GGML_SCHED_MAX_SPLIT_INPUTS) { + const size_t id = hash_id(src); + int src_backend_id = sched->tensor_backend_id[id]; + if (src_backend_id != cur_backend_id && sched->tensor_copies[hash_id(src)][cur_backend_id][0] == NULL) { + //printf("starting new split because of too many inputs: node %s, input %s\n", node->name, src->name); + need_new_split = true; + break; + } + } + } } - ggml_tallocr_t src_allocr = node_allocr(src); - if (src_allocr != node_allocr) { - int n_inputs = sched->splits[cur_split].n_inputs++; - GGML_ASSERT(n_inputs < GGML_MAX_SPLIT_INPUTS); - sched->splits[cur_split].inputs[n_inputs] = (struct ggml_tensor *)src; - - // create copies - size_t id = hash_id(src); - if (sched->node_copies[id][cur_backend_id] == NULL) { - struct ggml_tensor * tensor_copy = ggml_dup_tensor_layout(sched->ctx, src); - sched->node_copies[id][cur_backend_id] = tensor_copy; - node_allocr(tensor_copy) = cur_allocr; - ggml_backend_t backend = ggml_tallocr_get_buffer(cur_allocr)->backend; - ggml_format_name(tensor_copy, "%s#%s", ggml_backend_name(backend), src->name); + + if (node_backend_id != cur_backend_id || need_new_split) { + split->i_end = i; + i_split++; + if (i_split >= sched->splits_capacity) { + sched->splits_capacity *= 2; + sched->splits = realloc(sched->splits, sched->splits_capacity * sizeof(struct ggml_backend_sched_split)); + GGML_ASSERT(sched->splits != NULL); } - node->src[j] = sched->node_copies[id][cur_backend_id]; + GGML_ASSERT(i_split < GGML_SCHED_MAX_SPLITS); + split = &sched->splits[i_split]; + split->backend_id = node_backend_id; + split->i_start = i; + split->n_inputs = 0; + cur_backend_id = node_backend_id; } - } - } - sched->splits[cur_split].i_end = graph->n_nodes; - sched->n_splits = cur_split + 1; - //fprintf(stderr, "PASS 4 ASSIGNMENTS\n"); sched_print_assignments(sched, graph); fflush(stdout); + // find inputs that are not on the same backend + for (int j = 0; j < GGML_MAX_SRC; j++) { + struct ggml_tensor * src = node->src[j]; + if (src == NULL) { + continue; + } -#if 1 - // sanity check: all sources should have the same backend as the node - for (int i = 0; i < graph->n_nodes; i++) { - struct ggml_tensor * node = graph->nodes[i]; - ggml_tallocr_t node_allocr = node_allocr(node); - if (node_allocr == NULL) { - fprintf(stderr, "!!!!!!! %s has no backend\n", node->name); - } - for (int j = 0; j < GGML_MAX_SRC; j++) { - struct ggml_tensor * src = node->src[j]; - if (src == NULL) { - break; - } - ggml_tallocr_t src_allocr = node_allocr(src); - if (src_allocr != node_allocr /* && src_backend != NULL */) { // ignore nulls for now - fprintf(stderr, "!!!! %s has backend %s, src %d (%s) has backend %s\n", - node->name, node_allocr ? ggml_backend_name(ggml_tallocr_get_buffer(node_allocr)->backend) : "NULL", - j, src->name, src_allocr ? ggml_backend_name(ggml_tallocr_get_buffer(src_allocr)->backend) : "NULL"); + const int src_backend_id = tensor_backend_id(src); + assert(src_backend_id != -1); // all inputs should be assigned by now + + if (src->flags & GGML_TENSOR_FLAG_INPUT && sched->n_copies > 1) { + size_t id = hash_id(src); + if (sched->tensor_copies[id][src_backend_id][0] == NULL) { + ggml_backend_t backend = sched->backends[src_backend_id]; + for (int c = 0; c < sched->n_copies; c++) { + struct ggml_tensor * tensor_copy; + if (c == sched->cur_copy) { + tensor_copy = src; // use the original tensor as the current copy + } else { + tensor_copy = ggml_dup_tensor_layout(sched->ctx, src); + ggml_format_name(tensor_copy, "%s#%s#%d", ggml_backend_name(backend), src->name, c); + } + if (sched->n_copies > 1) { + ggml_set_input(tensor_copy); + ggml_set_output(tensor_copy); // prevent ggml-alloc from overwriting the tensor + } + sched->tensor_copies[id][src_backend_id][c] = tensor_copy; + SET_CAUSE(tensor_copy, "4.cpy"); + } + int n_graph_inputs = sched->n_graph_inputs++; + GGML_ASSERT(n_graph_inputs < GGML_SCHED_MAX_SPLIT_INPUTS); + sched->graph_inputs[n_graph_inputs] = src; + } + } + + if (src_backend_id != node_backend_id) { + // create a copy of the input in the split's backend + const size_t id = hash_id(src); + if (sched->tensor_copies[id][cur_backend_id][0] == NULL) { + ggml_backend_t backend = sched->backends[cur_backend_id]; + for (int c = 0; c < sched->n_copies; c++) { + struct ggml_tensor * tensor_copy = ggml_dup_tensor_layout(sched->ctx, src); + ggml_format_name(tensor_copy, "%s#%s#%d", ggml_backend_name(backend), src->name, c); + if (sched->n_copies > 1) { + ggml_set_input(tensor_copy); + ggml_set_output(tensor_copy); // prevent ggml-alloc from overwriting the tensor + } + sched->tensor_copies[id][cur_backend_id][c] = tensor_copy; + SET_CAUSE(tensor_copy, "4.cpy"); + } + int n_inputs = split->n_inputs++; + GGML_ASSERT(n_inputs < GGML_SCHED_MAX_SPLIT_INPUTS); + split->inputs[n_inputs] = src; + } + node->src[j] = sched->tensor_copies[id][cur_backend_id][sched->cur_copy]; + } } } + split->i_end = graph->n_nodes; + sched->n_splits = i_split + 1; } +#ifdef DEBUG_PASS4 + fprintf(stderr, "PASS 4 ASSIGNMENTS\n"); ggml_backend_sched_print_assignments(sched, graph); #endif // create copies of the graph for each split - // FIXME: avoid this copy, pass split inputs to ggml_gallocr_alloc_graph_n in some other way - struct ggml_cgraph * graph_copy = ggml_new_graph_custom(sched->ctx, graph->n_nodes + sched->n_splits*GGML_MAX_SPLIT_INPUTS, false); + // TODO: avoid this copy + struct ggml_cgraph * graph_copy = ggml_new_graph_custom(sched->ctx, graph->n_nodes + sched->n_splits*GGML_SCHED_MAX_SPLIT_INPUTS*2, false); for (int i = 0; i < sched->n_splits; i++) { struct ggml_backend_sched_split * split = &sched->splits[i]; - split->graph = ggml_graph_view(sched->ctx, graph, split->i_start, split->i_end); + split->graph = ggml_graph_view(graph, split->i_start, split->i_end); // add inputs to the graph copy so that they are allocated by ggml-alloc at the start of the split for (int j = 0; j < split->n_inputs; j++) { + assert(graph_copy->size > (graph_copy->n_nodes + 1)); + struct ggml_tensor * input = split->inputs[j]; - struct ggml_tensor * input_cpy = sched->node_copies[hash_id(input)][sched_allocr_prio(sched, split->tallocr)]; - input_cpy->src[0] = input; + const size_t input_id = hash_id(input); + struct ggml_tensor * input_cpy = sched->tensor_copies[input_id][split->backend_id][sched->cur_copy]; + + // add a dependency to the input source so that it is not freed before the copy is done + struct ggml_tensor * input_dep = ggml_view_tensor(sched->ctx, input); + input_dep->src[0] = input; + sched->node_backend_ids[graph_copy->n_nodes] = sched->tensor_backend_id[input_id]; + graph_copy->nodes[graph_copy->n_nodes++] = input_dep; + + // add a dependency to the input copy so that it is allocated at the start of the split + sched->node_backend_ids[graph_copy->n_nodes] = split->backend_id; graph_copy->nodes[graph_copy->n_nodes++] = input_cpy; } for (int j = split->i_start; j < split->i_end; j++) { + assert(graph_copy->size > graph_copy->n_nodes); + sched->node_backend_ids[graph_copy->n_nodes] = tensor_backend_id(graph->nodes[j]); graph_copy->nodes[graph_copy->n_nodes++] = graph->nodes[j]; } } + + if (sched->n_copies > 1) { + // add input copies as leafs so that they are allocated first + for (int i = 0; i < sched->n_graph_inputs; i++) { + struct ggml_tensor * input = sched->graph_inputs[i]; + size_t id = hash_id(input); + int backend_id = tensor_backend_id(input); + for (int c = 0; c < sched->n_copies; c++) { + struct ggml_tensor * input_cpy = sched->tensor_copies[id][backend_id][c]; + sched->leaf_backend_ids[graph_copy->n_leafs] = backend_id; + graph_copy->leafs[graph_copy->n_leafs++] = input_cpy; + } + } + + for (int i = 0; i < sched->n_splits; i++) { + struct ggml_backend_sched_split * split = &sched->splits[i]; + int backend_id = split->backend_id; + for (int j = 0; j < split->n_inputs; j++) { + struct ggml_tensor * input = split->inputs[j]; + size_t id = hash_id(input); + for (int c = 0; c < sched->n_copies; c++) { + struct ggml_tensor * input_cpy = sched->tensor_copies[id][backend_id][c]; + sched->leaf_backend_ids[graph_copy->n_leafs] = backend_id; + graph_copy->leafs[graph_copy->n_leafs++] = input_cpy; + } + } + } + } + + // add leafs from the original graph + for (int i = 0; i < graph->n_leafs; i++) { + struct ggml_tensor * leaf = graph->leafs[i]; + sched->leaf_backend_ids[graph_copy->n_leafs] = tensor_backend_id(leaf); + graph_copy->leafs[graph_copy->n_leafs++] = leaf; + } + sched->graph = graph_copy; } -static void sched_alloc_splits(ggml_backend_sched_t sched) { - ggml_gallocr_alloc_graph_n( - sched->galloc, - sched->graph, - sched->hash_set, - sched->node_talloc); -} +static bool ggml_backend_sched_alloc_splits(ggml_backend_sched_t sched) { + // allocate graph + if (!ggml_gallocr_alloc_graph(sched->galloc, sched->graph)) { + // the re-allocation may cause the split inputs to be moved to a different address + ggml_backend_sched_synchronize(sched); +#ifndef NDEBUG + fprintf(stderr, "%s: failed to allocate graph, reserving\n", __func__); +#endif + ggml_gallocr_reserve_n(sched->galloc, sched->graph, sched->node_backend_ids, sched->leaf_backend_ids); + if (!ggml_gallocr_alloc_graph(sched->galloc, sched->graph)) { + fprintf(stderr, "%s: failed to allocate graph\n", __func__); + return false; + } + } -static void sched_compute_splits(ggml_backend_sched_t sched) { - uint64_t copy_us[GGML_MAX_BACKENDS] = {0}; - uint64_t compute_us[GGML_MAX_BACKENDS] = {0}; + return true; +} +static enum ggml_status ggml_backend_sched_compute_splits(ggml_backend_sched_t sched) { struct ggml_backend_sched_split * splits = sched->splits; for (int i = 0; i < sched->n_splits; i++) { struct ggml_backend_sched_split * split = &splits[i]; - ggml_backend_t split_backend = ggml_tallocr_get_buffer(split->tallocr)->backend; - int split_backend_id = sched_backend_prio(sched, split_backend); + int split_backend_id = split->backend_id; + ggml_backend_t split_backend = sched->backends[split_backend_id]; // copy the input tensors to the split backend - uint64_t copy_start_us = ggml_time_us(); for (int j = 0; j < split->n_inputs; j++) { - struct ggml_tensor * input_cpy = sched->node_copies[hash_id(split->inputs[j])][sched_backend_prio(sched, split_backend)]; - if (split->inputs[j]->buffer == NULL) { - if (split->inputs[j]->view_src == NULL) { - fprintf(stderr, "input %s has no buffer and no view_src\n", split->inputs[j]->name); - exit(1); + ggml_backend_t input_backend = ggml_backend_sched_get_tensor_backend(sched, split->inputs[j]); + struct ggml_tensor * input = split->inputs[j]; + struct ggml_tensor * input_cpy = sched->tensor_copies[hash_id(input)][split_backend_id][sched->cur_copy]; + + if (input->flags & GGML_TENSOR_FLAG_INPUT) { + // inputs from the user must be copied immediately to prevent the user overwriting the data before the copy is done + if (sched->events[split_backend_id][sched->cur_copy] != NULL) { + ggml_backend_event_synchronize(sched->events[split_backend_id][sched->cur_copy]); + } else { + ggml_backend_synchronize(split_backend); } - struct ggml_tensor * view = split->inputs[j]; - view->backend = view->view_src->backend; - view->buffer = view->view_src->buffer; - view->data = (char *)view->view_src->data + view->view_offs; - ggml_backend_buffer_init_tensor(ggml_backend_sched_get_buffer(sched, view->buffer->backend), view); - } - if (input_cpy->buffer == NULL) { - fprintf(stderr, "input_cpy %s has no buffer\n", input_cpy->name); - exit(1); + ggml_backend_tensor_copy(input, input_cpy); + } else { + // wait for the split backend to finish using the input before overwriting it + if (sched->events[split_backend_id][sched->cur_copy] != NULL) { + ggml_backend_event_wait(split_backend, sched->events[split_backend_id][sched->cur_copy]); + } else { + ggml_backend_synchronize(split_backend); + } + ggml_backend_tensor_copy_async(input_backend, split_backend, input, input_cpy); } - GGML_ASSERT(split->inputs[j]->buffer->backend != input_cpy->buffer->backend); - GGML_ASSERT(input_cpy->buffer->backend == split_backend); - ggml_backend_tensor_copy(split->inputs[j], input_cpy); } - // ggml_backend_synchronize(split_backend); - int64_t copy_end_us = ggml_time_us(); - copy_us[split_backend_id] += copy_end_us - copy_start_us; -#if 0 - char split_filename[GGML_MAX_NAME]; - snprintf(split_filename, GGML_MAX_NAME, "split_%i_%s.dot", i, ggml_backend_name(split_backend)); - ggml_graph_dump_dot(split->graph, NULL, split_filename); -#endif + if (!sched->callback_eval) { + enum ggml_status ec = ggml_backend_graph_compute_async(split_backend, &split->graph); + if (ec != GGML_STATUS_SUCCESS) { + return ec; + } + } else { + // similar to ggml_backend_compare_graph_backend + for (int j0 = 0; j0 < split->graph.n_nodes; j0++) { + struct ggml_tensor * t = split->graph.nodes[j0]; - uint64_t compute_start_us = ggml_time_us(); - ggml_backend_graph_compute(split_backend, split->graph); - // ggml_backend_synchronize(split_backend); - uint64_t compute_end_us = ggml_time_us(); - compute_us[split_backend_id] += compute_end_us - compute_start_us; - } + // check if the user needs data from this node + bool need = sched->callback_eval(t, true, sched->callback_eval_user_data); -#if 0 - // per-backend timings - fprintf(stderr, "sched_compute_splits times (%d splits):\n", sched->n_splits); - for (int i = 0; i < sched->n_backends; i++) { - if (copy_us[i] > 0 || compute_us[i] > 0) { - fprintf(stderr, "\t%5.5s: %lu us copy, %lu us compute\n", ggml_backend_name(sched->backends[i]), copy_us[i], compute_us[i]); + int j1 = j0; + + // determine the range [j0, j1] of nodes that can be computed together + while (!need && j1 < split->graph.n_nodes - 1) { + t = split->graph.nodes[++j1]; + need = sched->callback_eval(t, true, sched->callback_eval_user_data); + } + + struct ggml_cgraph gv = ggml_graph_view(&split->graph, j0, j1 + 1); + + enum ggml_status ec = ggml_backend_graph_compute_async(split_backend, &gv); + if (ec != GGML_STATUS_SUCCESS) { + return ec; + } + + // TODO: pass backend to the callback, then the user can decide if they want to synchronize + ggml_backend_synchronize(split_backend); + + if (need && !sched->callback_eval(t, false, sched->callback_eval_user_data)) { + break; + } + + j0 = j1; + } } - } -#endif -} -static void sched_reset(ggml_backend_sched_t sched) { - for (int i = 0; i < sched->n_backends; i++) { - ggml_tallocr_reset(sched->tallocs[i]); + // record the event of this copy + if (split->n_inputs > 0) { + if (sched->events[split_backend_id][sched->cur_copy] != NULL) { + ggml_backend_event_record(sched->events[split_backend_id][sched->cur_copy]); + } + } } + + sched->cur_copy = (sched->cur_copy + 1) % sched->n_copies; + + return GGML_STATUS_SUCCESS; } -ggml_backend_sched_t ggml_backend_sched_new(ggml_backend_t * backends, int n_backends) { - GGML_ASSERT(n_backends <= GGML_MAX_BACKENDS); +ggml_backend_sched_t ggml_backend_sched_new( + ggml_backend_t * backends, + ggml_backend_buffer_type_t * bufts, + int n_backends, + size_t graph_size, + bool parallel) { + GGML_ASSERT(n_backends > 0); + GGML_ASSERT(n_backends <= GGML_SCHED_MAX_BACKENDS); + GGML_ASSERT(ggml_backend_is_cpu(backends[n_backends - 1])); // last backend must be CPU + + struct ggml_backend_sched * sched = calloc(sizeof(struct ggml_backend_sched), 1); - struct ggml_backend_sched * sched = malloc(sizeof(struct ggml_backend_sched)); - memset(sched, 0, sizeof(struct ggml_backend_sched)); + // initialize hash table + sched->hash_set = ggml_hash_set_new(graph_size); + sched->tensor_backend_id = calloc(sizeof(sched->tensor_backend_id[0]), sched->hash_set.size); + sched->tensor_copies = calloc(sizeof(sched->tensor_copies[0]), sched->hash_set.size); - fprintf(stderr, "ggml_backend_sched size: %lu KB\n", sizeof(struct ggml_backend_sched)/1024); + const size_t nodes_size = graph_size + GGML_SCHED_MAX_SPLITS*GGML_SCHED_MAX_SPLIT_INPUTS*2; + sched->node_backend_ids = calloc(sizeof(sched->node_backend_ids[0]), nodes_size); + sched->leaf_backend_ids = calloc(sizeof(sched->leaf_backend_ids[0]), nodes_size); sched->n_backends = n_backends; - for (int i = 0; i < n_backends; i++) { - sched->backends[i] = backends[i]; - } - sched->galloc = ggml_gallocr_new(); + sched->n_copies = parallel ? GGML_SCHED_MAX_COPIES : 1; - // init measure allocs for each backend - for (int i = 0; i < n_backends; i++) { - sched->tallocs[i] = ggml_tallocr_new_measure_from_backend(backends[i]); + const int initial_splits_capacity = 16; + sched->splits = calloc(sizeof(sched->splits[0]), initial_splits_capacity); + sched->splits_capacity = initial_splits_capacity; + + for (int b = 0; b < n_backends; b++) { + sched->backends[b] = backends[b]; + sched->bufts[b] = bufts ? bufts[b] : ggml_backend_get_default_buffer_type(backends[b]); + GGML_ASSERT(ggml_backend_buft_supports_backend(sched->bufts[b], backends[b])); + if (sched->n_copies > 1) { + for (int c = 0; c < sched->n_copies; c++) { + sched->events[b][c] = ggml_backend_event_new(backends[b]); + } + } } + sched->galloc = ggml_gallocr_new_n(sched->bufts, n_backends); + + ggml_backend_sched_reset(sched); + return sched; } @@ -893,58 +1762,334 @@ void ggml_backend_sched_free(ggml_backend_sched_t sched) { if (sched == NULL) { return; } - for (int i = 0; i < sched->n_backends; i++) { - ggml_tallocr_free(sched->tallocs[i]); + for (int b = 0; b < sched->n_backends; b++) { + for (int c = 0; c < sched->n_copies; c++) { + ggml_backend_event_free(sched->events[b][c]); + } } ggml_gallocr_free(sched->galloc); + ggml_free(sched->ctx); + free(sched->splits); free(sched->hash_set.keys); - free(sched->node_talloc); - free(sched->node_copies); + free(sched->tensor_backend_id); + free(sched->tensor_copies); + free(sched->node_backend_ids); + free(sched->leaf_backend_ids); free(sched); } -void ggml_backend_sched_init_measure(ggml_backend_sched_t sched, struct ggml_cgraph * measure_graph) { - // initialize hash tables - size_t hash_size = measure_graph->visited_hash_table.size + GGML_MAX_SPLITS*GGML_MAX_SPLIT_INPUTS; - sched->hash_set.size = hash_size; - sched->hash_set.keys = malloc(sizeof(sched->hash_set.keys[0]) * hash_size); - sched->node_talloc = malloc(sizeof(sched->node_talloc[0]) * hash_size); - sched->node_copies = malloc(sizeof(sched->node_copies[0]) * hash_size); +void ggml_backend_sched_reset(ggml_backend_sched_t sched) { + // reset state for the next run + size_t hash_size = sched->hash_set.size; + memset(sched->hash_set.keys, 0, sizeof(sched->hash_set.keys[0]) * hash_size); // NOLINT + memset(sched->tensor_backend_id, -1, sizeof(sched->tensor_backend_id[0]) * hash_size); + memset(sched->tensor_copies, 0, sizeof(sched->tensor_copies[0]) * hash_size); + + sched->is_reset = true; + sched->is_alloc = false; +} + +bool ggml_backend_sched_reserve(ggml_backend_sched_t sched, struct ggml_cgraph * measure_graph) { + GGML_ASSERT((int)sched->hash_set.size >= measure_graph->n_nodes); - sched_split_graph(sched, measure_graph); - sched_alloc_splits(sched); + ggml_backend_sched_split_graph(sched, measure_graph); - // allocate buffers and reset allocators - for (int i = 0; i < sched->n_backends; i++) { - size_t size = ggml_tallocr_max_size(sched->tallocs[i]); - ggml_tallocr_free(sched->tallocs[i]); - sched->tallocs[i] = ggml_tallocr_new_from_backend(sched->backends[i], size); + // TODO: extract this to a separate function + if (!ggml_gallocr_reserve_n(sched->galloc, sched->graph, sched->node_backend_ids, sched->leaf_backend_ids)) { + return false; + } + + ggml_backend_sched_reset(sched); + ggml_backend_sched_synchronize(sched); + + return true; +} + +bool ggml_backend_sched_alloc_graph(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { + GGML_ASSERT((int)sched->hash_set.size >= graph->n_nodes); + + ggml_backend_sched_split_graph(sched, graph); + + if (!ggml_backend_sched_alloc_splits(sched)) { + return false; + } + + sched->is_alloc = true; + + return true; +} + +enum ggml_status ggml_backend_sched_graph_compute(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { + enum ggml_status err = ggml_backend_sched_graph_compute_async(sched, graph); + ggml_backend_sched_synchronize(sched); + return err; +} + +enum ggml_status ggml_backend_sched_graph_compute_async(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { + if (!sched->is_reset && !sched->is_alloc) { + ggml_backend_sched_reset(sched); + } + + if (!sched->is_alloc) { + if (!ggml_backend_sched_alloc_graph(sched, graph)) { + return GGML_STATUS_ALLOC_FAILED; + } } - sched_reset(sched); + return ggml_backend_sched_compute_splits(sched); +} + +void ggml_backend_sched_synchronize(ggml_backend_sched_t sched) { + for (int i = 0; i < sched->n_backends; i++) { + ggml_backend_synchronize(sched->backends[i]); + } } -void ggml_backend_sched_graph_compute(ggml_backend_sched_t sched, struct ggml_cgraph * graph) { - GGML_ASSERT(sched->hash_set.size >= graph->visited_hash_table.size + GGML_MAX_SPLITS*GGML_MAX_SPLIT_INPUTS); +void ggml_backend_sched_set_eval_callback(ggml_backend_sched_t sched, ggml_backend_sched_eval_callback callback, void * user_data) { + sched->callback_eval = callback; + sched->callback_eval_user_data = user_data; +} - sched_split_graph(sched, graph); - sched_alloc_splits(sched); - sched_compute_splits(sched); - sched_reset(sched); +int ggml_backend_sched_get_n_splits(ggml_backend_sched_t sched) { + return sched->n_splits; } -ggml_tallocr_t ggml_backend_sched_get_tallocr(ggml_backend_sched_t sched, ggml_backend_t backend) { - int backend_index = sched_backend_prio(sched, backend); - return sched->tallocs[backend_index]; +int ggml_backend_sched_get_n_copies(ggml_backend_sched_t sched) { + return sched->n_copies; } -ggml_backend_buffer_t ggml_backend_sched_get_buffer(ggml_backend_sched_t sched, ggml_backend_t backend) { - int backend_index = sched_backend_prio(sched, backend); - return ggml_tallocr_get_buffer(sched->tallocs[backend_index]); +size_t ggml_backend_sched_get_buffer_size(ggml_backend_sched_t sched, ggml_backend_t backend) { + int backend_index = ggml_backend_sched_backend_id(sched, backend); + GGML_ASSERT(backend_index >= 0 && backend_index < sched->n_backends); + + return ggml_gallocr_get_buffer_size(sched->galloc, backend_index); } -void ggml_backend_sched_set_node_backend(ggml_backend_sched_t sched, struct ggml_tensor * node, ggml_backend_t backend) { - int backend_index = sched_backend_prio(sched, backend); +void ggml_backend_sched_set_tensor_backend(ggml_backend_sched_t sched, struct ggml_tensor * node, ggml_backend_t backend) { + int backend_index = ggml_backend_sched_backend_id(sched, backend); GGML_ASSERT(backend_index >= 0 && backend_index < sched->n_backends); - node_allocr(node) = sched->tallocs[backend_index]; + tensor_backend_id(node) = backend_index; +} + +ggml_backend_t ggml_backend_sched_get_tensor_backend(ggml_backend_sched_t sched, struct ggml_tensor * node) { + int backend_index = tensor_backend_id(node); + if (backend_index == -1) { + return NULL; + } + return sched->backends[backend_index]; +} + +// utils + +void ggml_backend_view_init(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor) { + GGML_ASSERT(tensor->buffer == NULL); + GGML_ASSERT(tensor->view_src != NULL); + GGML_ASSERT(tensor->view_src->buffer != NULL); + GGML_ASSERT(tensor->view_src->data != NULL); + + tensor->buffer = buffer; + tensor->data = (char *)tensor->view_src->data + tensor->view_offs; + tensor->backend = tensor->view_src->backend; + ggml_backend_buffer_init_tensor(buffer, tensor); +} + +void ggml_backend_tensor_alloc(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, void * addr) { + GGML_ASSERT(tensor->buffer == NULL); + GGML_ASSERT(tensor->data == NULL); + GGML_ASSERT(tensor->view_src == NULL); + GGML_ASSERT(addr >= ggml_backend_buffer_get_base(buffer)); + GGML_ASSERT((char *)addr + ggml_backend_buffer_get_alloc_size(buffer, tensor) <= + (char *)ggml_backend_buffer_get_base(buffer) + ggml_backend_buffer_get_size(buffer)); + + tensor->buffer = buffer; + tensor->data = addr; + ggml_backend_buffer_init_tensor(buffer, tensor); +} + +static struct ggml_tensor * graph_copy_dup_tensor(struct ggml_hash_set hash_set, struct ggml_tensor ** node_copies, + struct ggml_context * ctx_allocated, struct ggml_context * ctx_unallocated, struct ggml_tensor * src) { + + GGML_ASSERT(src != NULL); + GGML_ASSERT(src->data && "graph must be allocated"); + + size_t id = ggml_hash_insert(hash_set, src); + if (id == GGML_HASHTABLE_ALREADY_EXISTS) { + return node_copies[ggml_hash_find(hash_set, src)]; + } + + struct ggml_tensor * dst = ggml_dup_tensor_layout(src->data && !src->view_src ? ctx_allocated : ctx_unallocated, src); + if (src->view_src != NULL) { + dst->view_src = graph_copy_dup_tensor(hash_set, node_copies, ctx_allocated, ctx_unallocated, src->view_src); + dst->view_offs = src->view_offs; + } + dst->op = src->op; + memcpy(dst->op_params, src->op_params, sizeof(dst->op_params)); + ggml_set_name(dst, src->name); + + // copy src + for (int i = 0; i < GGML_MAX_SRC; i++) { + struct ggml_tensor * s = src->src[i]; + if (s == NULL) { + continue; + } + dst->src[i] = graph_copy_dup_tensor(hash_set, node_copies, ctx_allocated, ctx_unallocated, s); + } + + node_copies[id] = dst; + return dst; +} + +static void graph_copy_init_tensor(struct ggml_hash_set hash_set, struct ggml_tensor ** node_copies, bool * node_init, struct ggml_tensor * src) { + size_t id = ggml_hash_find(hash_set, src); + if (node_init[id]) { + return; + } + node_init[id] = true; + + struct ggml_tensor * dst = node_copies[id]; + if (dst->view_src != NULL) { + graph_copy_init_tensor(hash_set, node_copies, node_init, src->view_src); + ggml_backend_view_init(dst->view_src->buffer, dst); + } + else { + ggml_backend_tensor_copy(src, dst); + } + + // init src + for (int i = 0; i < GGML_MAX_SRC; i++) { + struct ggml_tensor * s = src->src[i]; + if (s == NULL) { + continue; + } + graph_copy_init_tensor(hash_set, node_copies, node_init, s); + } +} + +struct ggml_backend_graph_copy ggml_backend_graph_copy(ggml_backend_t backend, struct ggml_cgraph * graph) { + struct ggml_hash_set hash_set = { + /* .size = */ graph->visited_hash_table.size, + /* .keys = */ calloc(sizeof(hash_set.keys[0]), graph->visited_hash_table.size) // NOLINT + }; + struct ggml_tensor ** node_copies = calloc(sizeof(node_copies[0]), hash_set.size); // NOLINT + bool * node_init = calloc(sizeof(node_init[0]), hash_set.size); + + struct ggml_init_params params = { + /* .mem_size = */ ggml_tensor_overhead()*hash_set.size + ggml_graph_overhead_custom(graph->size, false), + /* .mem_buffer = */ NULL, + /* .no_alloc = */ true + }; + + struct ggml_context * ctx_allocated = ggml_init(params); + struct ggml_context * ctx_unallocated = ggml_init(params); + + if (ctx_allocated == NULL || ctx_unallocated == NULL) { + fprintf(stderr, "failed to allocate context for graph copy\n"); + free(hash_set.keys); + free(node_copies); + free(node_init); + ggml_free(ctx_allocated); + ggml_free(ctx_unallocated); + return (struct ggml_backend_graph_copy) { + /* .buffer = */ NULL, + /* .ctx_allocated = */ NULL, + /* .ctx_unallocated = */ NULL, + /* .graph = */ NULL, + }; + } + + // dup nodes + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + graph_copy_dup_tensor(hash_set, node_copies, ctx_allocated, ctx_unallocated, node); + } + + // allocate nodes + ggml_backend_buffer_t buffer = ggml_backend_alloc_ctx_tensors(ctx_allocated, backend); + if (buffer == NULL) { + fprintf(stderr, "failed to allocate buffer for graph copy\n"); + free(hash_set.keys); + free(node_copies); + free(node_init); + ggml_free(ctx_allocated); + ggml_free(ctx_unallocated); + return (struct ggml_backend_graph_copy) { + /* .buffer = */ NULL, + /* .ctx_allocated = */ NULL, + /* .ctx_unallocated = */ NULL, + /* .graph = */ NULL, + }; + } + + //printf("copy buffer size: %zu MB\n", ggml_backend_buffer_get_size(buffer) / 1024 / 1024); + + // copy data and init views + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + graph_copy_init_tensor(hash_set, node_copies, node_init, node); + } + + // build graph copy + struct ggml_cgraph * graph_copy = ggml_new_graph_custom(ctx_allocated, graph->size, false); + for (int i = 0; i < graph->n_nodes; i++) { + struct ggml_tensor * node = graph->nodes[i]; + struct ggml_tensor * node_copy = node_copies[ggml_hash_find(hash_set, node)]; + graph_copy->nodes[i] = node_copy; + } + graph_copy->n_nodes = graph->n_nodes; + + free(hash_set.keys); + free(node_copies); + free(node_init); + + return (struct ggml_backend_graph_copy) { + /* .buffer = */ buffer, + /* .ctx_allocated = */ ctx_allocated, + /* .ctx_unallocated = */ ctx_unallocated, + /* .graph = */ graph_copy, + }; +} + +void ggml_backend_graph_copy_free(struct ggml_backend_graph_copy copy) { + ggml_backend_buffer_free(copy.buffer); + ggml_free(copy.ctx_allocated); + ggml_free(copy.ctx_unallocated); +} + +bool ggml_backend_compare_graph_backend(ggml_backend_t backend1, ggml_backend_t backend2, struct ggml_cgraph * graph, ggml_backend_eval_callback callback, void * user_data) { + struct ggml_backend_graph_copy copy = ggml_backend_graph_copy(backend2, graph); + if (copy.buffer == NULL) { + return false; + } + + struct ggml_cgraph * g1 = graph; + struct ggml_cgraph * g2 = copy.graph; + + assert(g1->n_nodes == g2->n_nodes); + + for (int i = 0; i < g1->n_nodes; i++) { + //printf("eval %d/%d\n", i, g1->n_nodes); + struct ggml_tensor * t1 = g1->nodes[i]; + struct ggml_tensor * t2 = g2->nodes[i]; + + assert(t1->op == t2->op && ggml_are_same_layout(t1, t2)); + + struct ggml_cgraph g1v = ggml_graph_view(g1, i, i + 1); + struct ggml_cgraph g2v = ggml_graph_view(g2, i, i + 1); + + ggml_backend_graph_compute(backend1, &g1v); + ggml_backend_graph_compute(backend2, &g2v); + + if (ggml_is_view_op(t1->op)) { + continue; + } + + // compare results, calculate rms etc + if (!callback(i, t1, t2, user_data)) { + break; + } + } + + ggml_backend_graph_copy_free(copy); + + return true; } diff --git a/bindings/ruby/ext/ggml-backend.h b/bindings/ruby/ext/ggml-backend.h index 793a0a9d65a..744b6a77457 100644 --- a/bindings/ruby/ext/ggml-backend.h +++ b/bindings/ruby/ext/ggml-backend.h @@ -7,69 +7,123 @@ extern "C" { #endif + typedef struct ggml_backend_buffer_type * ggml_backend_buffer_type_t; + typedef struct ggml_backend_buffer * ggml_backend_buffer_t; + typedef struct ggml_backend_event * ggml_backend_event_t; + typedef struct ggml_backend * ggml_backend_t; + typedef void * ggml_backend_graph_plan_t; + // // Backend buffer // - struct ggml_backend_buffer; - typedef struct ggml_backend_buffer * ggml_backend_buffer_t; - - // backend buffer functions - GGML_API void ggml_backend_buffer_free (ggml_backend_buffer_t buffer); - GGML_API size_t ggml_backend_buffer_get_alignment (ggml_backend_buffer_t buffer); - GGML_API void * ggml_backend_buffer_get_base (ggml_backend_buffer_t buffer); - GGML_API size_t ggml_backend_buffer_get_size (ggml_backend_buffer_t buffer); - GGML_API size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); - GGML_API void ggml_backend_buffer_init_tensor (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); - GGML_API void ggml_backend_buffer_free_tensor (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); + // buffer type + GGML_API const char * ggml_backend_buft_name (ggml_backend_buffer_type_t buft); + GGML_API GGML_CALL ggml_backend_buffer_t ggml_backend_buft_alloc_buffer (ggml_backend_buffer_type_t buft, size_t size); + GGML_API size_t ggml_backend_buft_get_alignment (ggml_backend_buffer_type_t buft); + GGML_API size_t ggml_backend_buft_get_max_size (ggml_backend_buffer_type_t buft); + GGML_API GGML_CALL size_t ggml_backend_buft_get_alloc_size (ggml_backend_buffer_type_t buft, struct ggml_tensor * tensor); + GGML_API bool ggml_backend_buft_supports_backend(ggml_backend_buffer_type_t buft, ggml_backend_t backend); + GGML_API bool ggml_backend_buft_is_host (ggml_backend_buffer_type_t buft); + + // buffer + enum ggml_backend_buffer_usage { + GGML_BACKEND_BUFFER_USAGE_ANY = 0, + GGML_BACKEND_BUFFER_USAGE_WEIGHTS = 1, + }; + + GGML_API const char * ggml_backend_buffer_name (ggml_backend_buffer_t buffer); + GGML_API void ggml_backend_buffer_free (ggml_backend_buffer_t buffer); + GGML_API void * ggml_backend_buffer_get_base (ggml_backend_buffer_t buffer); + GGML_API size_t ggml_backend_buffer_get_size (ggml_backend_buffer_t buffer); + GGML_API GGML_CALL void ggml_backend_buffer_init_tensor (ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); + GGML_API size_t ggml_backend_buffer_get_alignment (ggml_backend_buffer_t buffer); + GGML_API size_t ggml_backend_buffer_get_max_size (ggml_backend_buffer_t buffer); + GGML_API size_t ggml_backend_buffer_get_alloc_size(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); + GGML_API void ggml_backend_buffer_clear (ggml_backend_buffer_t buffer, uint8_t value); + GGML_API bool ggml_backend_buffer_is_host (ggml_backend_buffer_t buffer); + GGML_API void ggml_backend_buffer_set_usage (ggml_backend_buffer_t buffer, enum ggml_backend_buffer_usage usage); + GGML_API ggml_backend_buffer_type_t ggml_backend_buffer_get_type (ggml_backend_buffer_t buffer); + GGML_API void ggml_backend_buffer_reset (ggml_backend_buffer_t buffer); // // Backend // - struct ggml_backend; - typedef struct ggml_backend * ggml_backend_t; - typedef void * ggml_backend_graph_plan_t; - - GGML_API ggml_backend_t ggml_get_backend(const struct ggml_tensor * tensor); - + GGML_API ggml_guid_t ggml_backend_guid(ggml_backend_t backend); GGML_API const char * ggml_backend_name(ggml_backend_t backend); GGML_API void ggml_backend_free(ggml_backend_t backend); - GGML_API ggml_backend_buffer_t ggml_backend_alloc_buffer(ggml_backend_t backend, size_t size); + GGML_API ggml_backend_buffer_type_t ggml_backend_get_default_buffer_type(ggml_backend_t backend); + GGML_API ggml_backend_buffer_t ggml_backend_alloc_buffer(ggml_backend_t backend, size_t size); + GGML_API size_t ggml_backend_get_alignment(ggml_backend_t backend); + GGML_API size_t ggml_backend_get_max_size(ggml_backend_t backend); - GGML_API size_t ggml_backend_get_alignment(ggml_backend_t backend); + GGML_API void ggml_backend_tensor_set_async(ggml_backend_t backend, struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); + GGML_API void ggml_backend_tensor_get_async(ggml_backend_t backend, const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); - GGML_API void ggml_backend_tensor_set_async( struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); - GGML_API void ggml_backend_tensor_get_async(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); - - GGML_API void ggml_backend_tensor_set( struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); - GGML_API void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); + GGML_API GGML_CALL void ggml_backend_tensor_set( struct ggml_tensor * tensor, const void * data, size_t offset, size_t size); + GGML_API GGML_CALL void ggml_backend_tensor_get(const struct ggml_tensor * tensor, void * data, size_t offset, size_t size); GGML_API void ggml_backend_synchronize(ggml_backend_t backend); - GGML_API ggml_backend_graph_plan_t ggml_backend_graph_plan_create (ggml_backend_t backend, struct ggml_cgraph * cgraph); + GGML_API ggml_backend_graph_plan_t ggml_backend_graph_plan_create(ggml_backend_t backend, struct ggml_cgraph * cgraph); + GGML_API void ggml_backend_graph_plan_free (ggml_backend_t backend, ggml_backend_graph_plan_t plan); - GGML_API void ggml_backend_graph_plan_free (ggml_backend_t backend, ggml_backend_graph_plan_t plan); - GGML_API void ggml_backend_graph_plan_compute(ggml_backend_t backend, ggml_backend_graph_plan_t plan); - GGML_API bool ggml_backend_graph_compute (ggml_backend_t backend, struct ggml_cgraph * cgraph); - GGML_API bool ggml_backend_supports_op (ggml_backend_t backend, const struct ggml_tensor * op); + GGML_API enum ggml_status ggml_backend_graph_plan_compute (ggml_backend_t backend, ggml_backend_graph_plan_t plan); + GGML_API enum ggml_status ggml_backend_graph_compute (ggml_backend_t backend, struct ggml_cgraph * cgraph); + GGML_API enum ggml_status ggml_backend_graph_compute_async(ggml_backend_t backend, struct ggml_cgraph * cgraph); + GGML_API bool ggml_backend_supports_op(ggml_backend_t backend, const struct ggml_tensor * op); + GGML_API bool ggml_backend_offload_op(ggml_backend_t backend, const struct ggml_tensor * op); // tensor copy between different backends GGML_API void ggml_backend_tensor_copy(struct ggml_tensor * src, struct ggml_tensor * dst); + // asynchronous copy + // the copy is performed after all the currently queued operations in backend_src + // backend_dst will wait for the copy to complete before performing other operations + // automatic fallback to sync copy if async is not supported + GGML_API void ggml_backend_tensor_copy_async(ggml_backend_t backend_src, ggml_backend_t backend_dst, struct ggml_tensor * src, struct ggml_tensor * dst); + + // events + GGML_API ggml_backend_event_t ggml_backend_event_new (ggml_backend_t backend); + GGML_API void ggml_backend_event_free (ggml_backend_event_t event); + GGML_API void ggml_backend_event_record (ggml_backend_event_t event); + GGML_API void ggml_backend_event_synchronize(ggml_backend_event_t event); + GGML_API void ggml_backend_event_wait (ggml_backend_t backend, ggml_backend_event_t event); // wait async on event + // // CPU backend // GGML_API ggml_backend_t ggml_backend_cpu_init(void); - GGML_API bool ggml_backend_is_cpu(ggml_backend_t backend); - GGML_API void ggml_backend_cpu_set_n_threads(ggml_backend_t backend_cpu, int n_threads); + GGML_API GGML_CALL bool ggml_backend_is_cpu (ggml_backend_t backend); + GGML_API void ggml_backend_cpu_set_n_threads (ggml_backend_t backend_cpu, int n_threads); + GGML_API void ggml_backend_cpu_set_abort_callback(ggml_backend_t backend_cpu, ggml_abort_callback abort_callback, void * abort_callback_data); // Create a backend buffer from an existing pointer - GGML_API ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(ggml_backend_t backend_cpu, void * ptr, size_t size); + GGML_API GGML_CALL ggml_backend_buffer_t ggml_backend_cpu_buffer_from_ptr(void * ptr, size_t size); + + GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_cpu_buffer_type(void); + +#ifdef GGML_USE_CPU_HBM + GGML_API ggml_backend_buffer_type_t ggml_backend_cpu_hbm_buffer_type(void); +#endif + + // + // Backend registry + // + + // The backend registry is a registry of all the available backends, and allows initializing backends in a generic way + GGML_API size_t ggml_backend_reg_get_count(void); + GGML_API size_t ggml_backend_reg_find_by_name(const char * name); + GGML_API ggml_backend_t ggml_backend_reg_init_backend_from_str(const char * backend_str); // str is name[:params] + GGML_API const char * ggml_backend_reg_get_name(size_t i); + GGML_API ggml_backend_t ggml_backend_reg_init_backend(size_t i, const char * params); // params is backend-specific + GGML_API ggml_backend_buffer_type_t ggml_backend_reg_get_default_buffer_type(size_t i); + GGML_API ggml_backend_buffer_t ggml_backend_reg_alloc_buffer(size_t i, size_t size); // // Backend scheduler @@ -83,53 +137,96 @@ extern "C" { /* Example usage: - sched = ggml_backend_sched_new({backend_gpu, backend_gpu2, backend_cpu}, num_backends); - // sched is initialized with measure allocators and cannot be used until allocated with a measure graph + // operations that use tensors allocated in a buffer with USAGE_WEIGHTS will be assigned + // preferrably to run on the same backend as the buffer + ggml_backend_buffer_set_usage(buf_weights, GGML_BACKEND_BUFFER_USAGE_WEIGHTS); - // initialize buffers from a measure graph - measure_graph = build_graph(sched); // use the allocr to allocate inputs as needed + sched = ggml_backend_sched_new({backend_gpu, backend_gpu2, backend_cpu}, NULL, num_backends, GGML_DEFAULT_GRAPH_SIZE, false); - // in build_graph: - build_graph(...) { - // allocating tensors in a specific backend (optional, recommended: pre-allocate inputs in a different buffer) - alloc_cpu = ggml_backend_sched_get_allocr(sched, backend_cpu); - ggml_allocr_alloc(alloc_cpu, tensor); + // initialize buffers from a max size graph (optional) + reserve_graph = build_graph(sched, max_batch_size); - // manually assigning nodes to a backend (optional, shouldn't be needed in most cases) - struct ggml_tensor * node = ggml_mul_mat(ctx, ...); - ggml_backend_sched_set_node_backend(sched, node, backend_gpu); - } + // manually assign nodes to a backend (optional, should not be needed in most cases) + struct ggml_tensor * node = ggml_mul_mat(ctx, ...); + ggml_backend_sched_set_tensor_backend(sched, node, backend_gpu); - // allocate backend buffers from measure graph - ggml_backend_sched_init_measure(sched, measure_graph); - - // the scheduler is now ready to compute graphs + ggml_backend_sched_reserve(sched, reserve_graph); // compute graph = build_graph(sched); ggml_backend_sched_graph_compute(sched, graph); + + // if there are graph inputs: + ggml_backend_sched_reset(sched); + ggml_backend_sched_alloc_graph(sched, graph); + ggml_backend_tensor_set(input_tensor, ...); + ggml_backend_sched_graph_compute(sched, graph); + } */ struct ggml_backend_sched; typedef struct ggml_backend_sched * ggml_backend_sched_t; - // Initialize a backend scheduler - GGML_API ggml_backend_sched_t ggml_backend_sched_new(ggml_backend_t * backends, int n_backends); + // when ask == true, the scheduler wants to know if the user wants to observe this node + // this allows the scheduler to batch nodes together in order to evaluate them in a single call + // + // when ask == false, the scheduler is passing the node tensor to the user for observation + // if the user returns false, the scheduler will cancel the graph compute + // + typedef bool (*ggml_backend_sched_eval_callback)(struct ggml_tensor * t, bool ask, void * user_data); - GGML_API void ggml_backend_sched_free(ggml_backend_sched_t sched); + // Initialize a backend scheduler + GGML_API ggml_backend_sched_t ggml_backend_sched_new(ggml_backend_t * backends, ggml_backend_buffer_type_t * bufts, int n_backends, size_t graph_size, bool parallel); + GGML_API void ggml_backend_sched_free(ggml_backend_sched_t sched); // Initialize backend buffers from a measure graph - GGML_API void ggml_backend_sched_init_measure(ggml_backend_sched_t sched, struct ggml_cgraph * measure_graph); + GGML_API bool ggml_backend_sched_reserve(ggml_backend_sched_t sched, struct ggml_cgraph * measure_graph); + + // Get the number of splits of the last graph + GGML_API int ggml_backend_sched_get_n_splits(ggml_backend_sched_t sched); + GGML_API int ggml_backend_sched_get_n_copies(ggml_backend_sched_t sched); + + GGML_API size_t ggml_backend_sched_get_buffer_size(ggml_backend_sched_t sched, ggml_backend_t backend); + + GGML_API void ggml_backend_sched_set_tensor_backend(ggml_backend_sched_t sched, struct ggml_tensor * node, ggml_backend_t backend); + GGML_API ggml_backend_t ggml_backend_sched_get_tensor_backend(ggml_backend_sched_t sched, struct ggml_tensor * node); + + // Allocate and compute graph on the backend scheduler + GGML_API bool ggml_backend_sched_alloc_graph(ggml_backend_sched_t sched, struct ggml_cgraph * graph); + GGML_API enum ggml_status ggml_backend_sched_graph_compute(ggml_backend_sched_t sched, struct ggml_cgraph * graph); + GGML_API enum ggml_status ggml_backend_sched_graph_compute_async(ggml_backend_sched_t sched, struct ggml_cgraph * graph); + GGML_API void ggml_backend_sched_synchronize(ggml_backend_sched_t sched); + + // Reset all assignments and allocators - must be called before changing the node backends + GGML_API void ggml_backend_sched_reset(ggml_backend_sched_t sched); + + // Set a callback to be called for each resulting node during graph compute + GGML_API void ggml_backend_sched_set_eval_callback(ggml_backend_sched_t sched, ggml_backend_sched_eval_callback callback, void * user_data); + + // + // Utils + // + + struct ggml_backend_graph_copy { + ggml_backend_buffer_t buffer; + struct ggml_context * ctx_allocated; + struct ggml_context * ctx_unallocated; + struct ggml_cgraph * graph; + }; + + // Copy a graph to a different backend + GGML_API struct ggml_backend_graph_copy ggml_backend_graph_copy(ggml_backend_t backend, struct ggml_cgraph * graph); + GGML_API void ggml_backend_graph_copy_free(struct ggml_backend_graph_copy copy); + + typedef bool (*GGML_CALL ggml_backend_eval_callback)(int node_index, struct ggml_tensor * t1, struct ggml_tensor * t2, void * user_data); - GGML_API ggml_tallocr_t ggml_backend_sched_get_tallocr(ggml_backend_sched_t sched, ggml_backend_t backend); - GGML_API ggml_backend_buffer_t ggml_backend_sched_get_buffer (ggml_backend_sched_t sched, ggml_backend_t backend); + // Compare the output of two backends + GGML_API bool ggml_backend_compare_graph_backend(ggml_backend_t backend1, ggml_backend_t backend2, struct ggml_cgraph * graph, ggml_backend_eval_callback callback, void * user_data); - GGML_API void ggml_backend_sched_set_node_backend(ggml_backend_sched_t sched, struct ggml_tensor * node, ggml_backend_t backend); + // Tensor initialization + GGML_API void ggml_backend_tensor_alloc(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor, void * addr); + GGML_API void ggml_backend_view_init(ggml_backend_buffer_t buffer, struct ggml_tensor * tensor); - // Allocate a graph on the backend scheduler - GGML_API void ggml_backend_sched_graph_compute( - ggml_backend_sched_t sched, - struct ggml_cgraph * graph); #ifdef __cplusplus } diff --git a/bindings/ruby/ext/ggml-common.h b/bindings/ruby/ext/ggml-common.h new file mode 100644 index 00000000000..43c7978a098 --- /dev/null +++ b/bindings/ruby/ext/ggml-common.h @@ -0,0 +1,1853 @@ +#ifndef GGML_COMMON_DECL + +#if defined(GGML_COMMON_DECL_C) +#include + +typedef uint16_t ggml_half; +typedef uint32_t ggml_half2; + +#define GGML_COMMON_AGGR + +#define GGML_COMMON_DECL +#elif defined(GGML_COMMON_DECL_METAL) +#include + +typedef half ggml_half; +typedef half2 ggml_half2; + +#define GGML_COMMON_AGGR + +#define GGML_COMMON_DECL +#elif defined(GGML_COMMON_DECL_CUDA) +#include +#include + +typedef half ggml_half; +typedef half2 ggml_half2; + +#define GGML_COMMON_AGGR data + +#define GGML_COMMON_DECL +#elif defined(GGML_COMMON_DECL_HIP) +#include +#include + +typedef half ggml_half; +typedef half2 ggml_half2; + +#define GGML_COMMON_AGGR data + +#define GGML_COMMON_DECL +#elif defined(GGML_COMMON_DECL_SYCL) +#include +#include + +typedef sycl::half ggml_half; +typedef sycl::half2 ggml_half2; + +#define GGML_COMMON_AGGR data + +#define GGML_COMMON_DECL +#endif + +#if defined(GGML_COMMON_DECL) + +#ifndef __cplusplus +#ifndef static_assert +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201100L) +#define static_assert(cond, msg) _Static_assert(cond, msg) +#else +#define static_assert(cond, msg) struct global_scope_noop_trick +#endif +#endif +#endif // __cplusplus + +// QK = number of values after dequantization +// QK_K = super-block size + +#ifdef GGML_QKK_64 +#define QK_K 64 +#define K_SCALE_SIZE 4 +#else +#define QK_K 256 +#define K_SCALE_SIZE 12 +#endif // GGML_QKK_64 + +#if defined(GGML_COMMON_DECL_CUDA) || defined(GGML_COMMON_DECL_HIP) || defined(GGML_COMMON_DECL_SYCL) +// QR = QK / number of values before dequantization +// QI = number of 32 bit integers before dequantization + +#define QI4_0 (QK4_0 / (4 * QR4_0)) +#define QR4_0 2 + +#define QI4_1 (QK4_1 / (4 * QR4_1)) +#define QR4_1 2 + +#define QI5_0 (QK5_0 / (4 * QR5_0)) +#define QR5_0 2 + +#define QI5_1 (QK5_1 / (4 * QR5_1)) +#define QR5_1 2 + +#define QI8_0 (QK8_0 / (4 * QR8_0)) +#define QR8_0 1 + +#define QI8_1 (QK8_1 / (4 * QR8_1)) +#define QR8_1 1 + +#define QI2_K (QK_K / (4*QR2_K)) +#define QR2_K 4 + +#define QI3_K (QK_K / (4*QR3_K)) +#define QR3_K 4 + +#define QI4_K (QK_K / (4*QR4_K)) +#define QR4_K 2 + +#define QI5_K (QK_K / (4*QR5_K)) +#define QR5_K 2 + +#define QI6_K (QK_K / (4*QR6_K)) +#define QR6_K 2 + +#define QI2_XXS (QK_K / (4*QR2_XXS)) +#define QR2_XXS 8 + +#define QI2_XS (QK_K / (4*QR2_XS)) +#define QR2_XS 8 + +#define QI2_S (QK_K / (4*QR2_S)) +#define QR2_S 8 + +#define QI3_XXS (QK_K / (4*QR3_XXS)) +#define QR3_XXS 8 + +#define QI3_XS (QK_K / (4*QR3_XS)) +#define QR3_XS 8 + +#define QI1_S (QK_K / (4*QR1_S)) +#define QR1_S 8 + +#define QI4_NL (QK4_NL / (4*QR4_NL)) +#define QR4_NL 2 + +#if QK_K == 64 +#define QI4_XS QI4_NL +#define QR4_XS QR4_NL +#else +#define QI4_XS (QK_K / (4*QR4_XS)) +#define QR4_XS 8 +#endif + +#endif // GGML_COMMON_DECL_CUDA || GGML_COMMON_DECL_HIP + +#define QK4_0 32 +typedef struct { + ggml_half d; // delta + uint8_t qs[QK4_0 / 2]; // nibbles / quants +} block_q4_0; +static_assert(sizeof(block_q4_0) == sizeof(ggml_half) + QK4_0 / 2, "wrong q4_0 block size/padding"); + +#define QK4_1 32 +typedef struct { + union { + struct { + ggml_half d; // delta + ggml_half m; // min + } GGML_COMMON_AGGR; + ggml_half2 dm; + }; + uint8_t qs[QK4_1 / 2]; // nibbles / quants +} block_q4_1; +static_assert(sizeof(block_q4_1) == 2 * sizeof(ggml_half) + QK4_1 / 2, "wrong q4_1 block size/padding"); + +#define QK5_0 32 +typedef struct { + ggml_half d; // delta + uint8_t qh[4]; // 5-th bit of quants + uint8_t qs[QK5_0 / 2]; // nibbles / quants +} block_q5_0; +static_assert(sizeof(block_q5_0) == sizeof(ggml_half) + sizeof(uint32_t) + QK5_0 / 2, "wrong q5_0 block size/padding"); + +#define QK5_1 32 +typedef struct { + union { + struct { + ggml_half d; // delta + ggml_half m; // min + } GGML_COMMON_AGGR; + ggml_half2 dm; + }; + uint8_t qh[4]; // 5-th bit of quants + uint8_t qs[QK5_1 / 2]; // nibbles / quants +} block_q5_1; +static_assert(sizeof(block_q5_1) == 2 * sizeof(ggml_half) + sizeof(uint32_t) + QK5_1 / 2, "wrong q5_1 block size/padding"); + +#define QK8_0 32 +typedef struct { + ggml_half d; // delta + int8_t qs[QK8_0]; // quants +} block_q8_0; +static_assert(sizeof(block_q8_0) == sizeof(ggml_half) + QK8_0, "wrong q8_0 block size/padding"); + +#define QK8_1 32 +typedef struct { + union { + struct { + ggml_half d; // delta + ggml_half s; // d * sum(qs[i]) + } GGML_COMMON_AGGR; + ggml_half2 ds; + }; + int8_t qs[QK8_1]; // quants +} block_q8_1; +static_assert(sizeof(block_q8_1) == 2*sizeof(ggml_half) + QK8_1, "wrong q8_1 block size/padding"); + +// +// Super-block quantization structures +// + +// 2-bit quantization +// weight is represented as x = a * q + b +// 16 blocks of 16 elements each +// Effectively 2.625 bits per weight +typedef struct { + uint8_t scales[QK_K/16]; // scales and mins, quantized with 4 bits + uint8_t qs[QK_K/4]; // quants + union { + struct { + ggml_half d; // super-block scale for quantized scales + ggml_half dmin; // super-block scale for quantized mins + } GGML_COMMON_AGGR; + ggml_half2 dm; + }; +} block_q2_K; +static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_half) + QK_K/16 + QK_K/4, "wrong q2_K block size/padding"); + +// 3-bit quantization +// weight is represented as x = a * q +// 16 blocks of 16 elements each +// Effectively 3.4375 bits per weight +#ifdef GGML_QKK_64 +typedef struct { + uint8_t hmask[QK_K/8]; // quants - high bit + uint8_t qs[QK_K/4]; // quants - low 2 bits + uint8_t scales[2]; + ggml_half d; // super-block scale +} block_q3_K; +static_assert(sizeof(block_q3_K) == sizeof(ggml_half) + QK_K / 4 + QK_K / 8 + 2, "wrong q3_K block size/padding"); +#else +typedef struct { + uint8_t hmask[QK_K/8]; // quants - high bit + uint8_t qs[QK_K/4]; // quants - low 2 bits + uint8_t scales[12]; // scales, quantized with 6 bits + ggml_half d; // super-block scale +} block_q3_K; +static_assert(sizeof(block_q3_K) == sizeof(ggml_half) + QK_K / 4 + QK_K / 8 + 12, "wrong q3_K block size/padding"); +#endif + +// 4-bit quantization +// 8 blocks of 32 elements each +// weight is represented as x = a * q + b +// Effectively 4.5 bits per weight +#ifdef GGML_QKK_64 +typedef struct { + ggml_half d[2]; // super-block scales/mins + uint8_t scales[2]; // 4-bit block scales/mins + uint8_t qs[QK_K/2]; // 4--bit quants +} block_q4_K; +static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_half) + QK_K/2 + 2, "wrong q4_K block size/padding"); +#else +typedef struct { + union { + struct { + ggml_half d; // super-block scale for quantized scales + ggml_half dmin; // super-block scale for quantized mins + } GGML_COMMON_AGGR; + ggml_half2 dm; + }; + uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits + uint8_t qs[QK_K/2]; // 4--bit quants +} block_q4_K; +static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_half) + K_SCALE_SIZE + QK_K/2, "wrong q4_K block size/padding"); +#endif + +// 5-bit quantization +// 8 blocks of 32 elements each +// weight is represented as x = a * q + b +// Effectively 5.5 bits per weight +#ifdef GGML_QKK_64 +typedef struct { + ggml_half d; // super-block scale + int8_t scales[QK_K/16]; // 8-bit block scales + uint8_t qh[QK_K/8]; // quants, high bit + uint8_t qs[QK_K/2]; // quants, low 4 bits +} block_q5_K; +static_assert(sizeof(block_q5_K) == sizeof(ggml_half) + QK_K/2 + QK_K/8 + QK_K/16, "wrong q5_K block size/padding"); +#else +typedef struct { + union { + struct { + ggml_half d; // super-block scale for quantized scales + ggml_half dmin; // super-block scale for quantized mins + } GGML_COMMON_AGGR; + ggml_half2 dm; + }; + uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits + uint8_t qh[QK_K/8]; // quants, high bit + uint8_t qs[QK_K/2]; // quants, low 4 bits +} block_q5_K; +static_assert(sizeof(block_q5_K) == 2*sizeof(ggml_half) + K_SCALE_SIZE + QK_K/2 + QK_K/8, "wrong q5_K block size/padding"); +#endif + +// 6-bit quantization +// weight is represented as x = a * q +// 16 blocks of 16 elements each +// Effectively 6.5625 bits per weight +typedef struct { + uint8_t ql[QK_K/2]; // quants, lower 4 bits + uint8_t qh[QK_K/4]; // quants, upper 2 bits + int8_t scales[QK_K/16]; // scales, quantized with 8 bits + ggml_half d; // super-block scale +} block_q6_K; +static_assert(sizeof(block_q6_K) == sizeof(ggml_half) + QK_K / 16 + 3*QK_K/4, "wrong q6_K block size/padding"); + +// This is only used for intermediate quantization and dot products +typedef struct { + float d; // delta + int8_t qs[QK_K]; // quants + int16_t bsums[QK_K/16]; // sum of quants in groups of 16 +} block_q8_K; +static_assert(sizeof(block_q8_K) == sizeof(float) + QK_K + QK_K/16*sizeof(int16_t), "wrong q8_K block size/padding"); + +// (Almost) "true" 2-bit quantization. +// Due to the need to use blocks as per ggml design, it ends up using +// 2.0625 bpw because of the 16-bit scale for each block of 256. +typedef struct { + ggml_half d; + uint16_t qs[QK_K/8]; +} block_iq2_xxs; +static_assert(sizeof(block_iq2_xxs) == sizeof(ggml_half) + QK_K/8*sizeof(uint16_t), "wrong iq2_xxs block size/padding"); + +// 2.3125 bpw quants +typedef struct { + ggml_half d; + uint16_t qs[QK_K/8]; + uint8_t scales[QK_K/32]; +} block_iq2_xs; +static_assert(sizeof(block_iq2_xs) == sizeof(ggml_half) + QK_K/8*sizeof(uint16_t) + QK_K/32, "wrong iq2_xs block size/padding"); + +// 2.5625 bpw quants +typedef struct { + ggml_half d; + uint8_t qs[QK_K/4]; + uint8_t qh[QK_K/32]; + uint8_t scales[QK_K/32]; +} block_iq2_s; +static_assert(sizeof(block_iq2_s) == sizeof(ggml_half) + QK_K/4 + QK_K/16, "wrong iq2_s block size/padding"); + +// (Almost) "true" 3-bit quantization. +// Due to the need to use blocks as per ggml design, it ends up using +// 3.0625 bpw because of the 16-bit scale for each block of 256. +typedef struct { + ggml_half d; + uint8_t qs[3*QK_K/8]; +} block_iq3_xxs; +static_assert(sizeof(block_iq3_xxs) == sizeof(ggml_half) + 3*(QK_K/8), "wrong iq3_xxs block size/padding"); + +// 3.4375 bpw +#if QK_K == 64 +#define IQ3S_N_SCALE 2 +#else +#define IQ3S_N_SCALE QK_K/64 +#endif +typedef struct { + ggml_half d; + uint8_t qs[QK_K/4]; + uint8_t qh[QK_K/32]; + uint8_t signs[QK_K/8]; + uint8_t scales[IQ3S_N_SCALE]; +} block_iq3_s; +static_assert(sizeof(block_iq3_s) == sizeof(ggml_half) + 13*(QK_K/32) + IQ3S_N_SCALE, "wrong iq3_s block size/padding"); + +typedef struct { + ggml_half d; + uint8_t qs[QK_K/8]; + uint16_t qh[QK_K/32]; +} block_iq1_s; +static_assert(sizeof(block_iq1_s) == sizeof(ggml_half) + QK_K/8 + QK_K/16, "wrong iq1_s block size/padding"); + +// 1.75 bpw +typedef struct { + uint8_t qs[QK_K/8]; // grid index, low 8 bits + uint8_t qh[QK_K/16]; // grid index, high 3 bits + grid shift bit (for two groups of 8) +#if QK_K == 64 + ggml_half d; +#endif + uint8_t scales[QK_K/32]; // 3-bit block scales (4-bit if QK_K == 64) +} block_iq1_m; +#if QK_K == 64 +static_assert(sizeof(block_iq1_m) == QK_K/8 + QK_K/16 + QK_K/32 + sizeof(ggml_half), "wrong iq1_m block size/padding"); +#else +static_assert(sizeof(block_iq1_m) == QK_K/8 + QK_K/16 + QK_K/32, "wrong iq1_m block size/padding"); +#endif + +// Used by IQ1_M quants +typedef union { + ggml_half f16; + uint16_t u16; +} iq1m_scale_t; + +// Non-linear quants +#define QK4_NL 32 +typedef struct { + ggml_half d; + uint8_t qs[QK4_NL/2]; +} block_iq4_nl; +static_assert(sizeof(block_iq4_nl) == sizeof(ggml_half) + QK4_NL/2, "wrong iq4_nl block size/padding"); + +#if QK_K == 64 +#define block_iq4_xs block_iq4_nl +#else +typedef struct { + ggml_half d; + uint16_t scales_h; + uint8_t scales_l[QK_K/64]; + uint8_t qs[QK_K/2]; +} block_iq4_xs; +static_assert(sizeof(block_iq4_xs) == sizeof(ggml_half) + sizeof(uint16_t) + QK_K/64 + QK_K/2, "wrong iq4_xs block size/padding"); +#endif + +#endif // GGML_COMMON_DECL +#endif // GGML_COMMON_DECL + +//////////////////////////////////////////////////////////////////////////////// + +#ifndef GGML_COMMON_IMPL + +#if defined(GGML_COMMON_IMPL_C) +#include + +#define GGML_TABLE_BEGIN(type, name, size) static const type name[size] = { +#define GGML_TABLE_END() }; + +#define GGML_COMMON_IMPL +#elif defined(GGML_COMMON_IMPL_METAL) +#include + +#define GGML_TABLE_BEGIN(type, name, size) static const constant type name[size] = { +#define GGML_TABLE_END() }; + +#define GGML_COMMON_IMPL +#elif defined(GGML_COMMON_IMPL_CUDA) || defined(GGML_COMMON_IMPL_HIP) +#include + +#define GGML_TABLE_BEGIN(type, name, size) static const __device__ type name[size] = { +#define GGML_TABLE_END() }; + +#define GGML_COMMON_IMPL +#elif defined(GGML_COMMON_IMPL_SYCL) + +#include + +#define GGML_TABLE_BEGIN(type, name, size) static const type name[size] = { +#define GGML_TABLE_END() }; + +#define GGML_COMMON_IMPL +#endif + +#if defined(GGML_COMMON_IMPL) + +GGML_TABLE_BEGIN(uint8_t, kmask_iq2xs, 8) + 1, 2, 4, 8, 16, 32, 64, 128 +GGML_TABLE_END() + +GGML_TABLE_BEGIN(uint8_t, ksigns_iq2xs, 128) + 0, 129, 130, 3, 132, 5, 6, 135, 136, 9, 10, 139, 12, 141, 142, 15, + 144, 17, 18, 147, 20, 149, 150, 23, 24, 153, 154, 27, 156, 29, 30, 159, + 160, 33, 34, 163, 36, 165, 166, 39, 40, 169, 170, 43, 172, 45, 46, 175, + 48, 177, 178, 51, 180, 53, 54, 183, 184, 57, 58, 187, 60, 189, 190, 63, + 192, 65, 66, 195, 68, 197, 198, 71, 72, 201, 202, 75, 204, 77, 78, 207, + 80, 209, 210, 83, 212, 85, 86, 215, 216, 89, 90, 219, 92, 221, 222, 95, + 96, 225, 226, 99, 228, 101, 102, 231, 232, 105, 106, 235, 108, 237, 238, 111, + 240, 113, 114, 243, 116, 245, 246, 119, 120, 249, 250, 123, 252, 125, 126, 255, +GGML_TABLE_END() + +//#if __CUDA_ARCH__ >= MIN_CC_DP4A // lowest compute capability for integer intrinsics +GGML_TABLE_BEGIN(uint64_t, ksigns64, 128) + 0x0000000000000000, 0xff000000000000ff, 0xff0000000000ff00, 0x000000000000ffff, + 0xff00000000ff0000, 0x0000000000ff00ff, 0x0000000000ffff00, 0xff00000000ffffff, + 0xff000000ff000000, 0x00000000ff0000ff, 0x00000000ff00ff00, 0xff000000ff00ffff, + 0x00000000ffff0000, 0xff000000ffff00ff, 0xff000000ffffff00, 0x00000000ffffffff, + 0xff0000ff00000000, 0x000000ff000000ff, 0x000000ff0000ff00, 0xff0000ff0000ffff, + 0x000000ff00ff0000, 0xff0000ff00ff00ff, 0xff0000ff00ffff00, 0x000000ff00ffffff, + 0x000000ffff000000, 0xff0000ffff0000ff, 0xff0000ffff00ff00, 0x000000ffff00ffff, + 0xff0000ffffff0000, 0x000000ffffff00ff, 0x000000ffffffff00, 0xff0000ffffffffff, + 0xff00ff0000000000, 0x0000ff00000000ff, 0x0000ff000000ff00, 0xff00ff000000ffff, + 0x0000ff0000ff0000, 0xff00ff0000ff00ff, 0xff00ff0000ffff00, 0x0000ff0000ffffff, + 0x0000ff00ff000000, 0xff00ff00ff0000ff, 0xff00ff00ff00ff00, 0x0000ff00ff00ffff, + 0xff00ff00ffff0000, 0x0000ff00ffff00ff, 0x0000ff00ffffff00, 0xff00ff00ffffffff, + 0x0000ffff00000000, 0xff00ffff000000ff, 0xff00ffff0000ff00, 0x0000ffff0000ffff, + 0xff00ffff00ff0000, 0x0000ffff00ff00ff, 0x0000ffff00ffff00, 0xff00ffff00ffffff, + 0xff00ffffff000000, 0x0000ffffff0000ff, 0x0000ffffff00ff00, 0xff00ffffff00ffff, + 0x0000ffffffff0000, 0xff00ffffffff00ff, 0xff00ffffffffff00, 0x0000ffffffffffff, + 0xffff000000000000, 0x00ff0000000000ff, 0x00ff00000000ff00, 0xffff00000000ffff, + 0x00ff000000ff0000, 0xffff000000ff00ff, 0xffff000000ffff00, 0x00ff000000ffffff, + 0x00ff0000ff000000, 0xffff0000ff0000ff, 0xffff0000ff00ff00, 0x00ff0000ff00ffff, + 0xffff0000ffff0000, 0x00ff0000ffff00ff, 0x00ff0000ffffff00, 0xffff0000ffffffff, + 0x00ff00ff00000000, 0xffff00ff000000ff, 0xffff00ff0000ff00, 0x00ff00ff0000ffff, + 0xffff00ff00ff0000, 0x00ff00ff00ff00ff, 0x00ff00ff00ffff00, 0xffff00ff00ffffff, + 0xffff00ffff000000, 0x00ff00ffff0000ff, 0x00ff00ffff00ff00, 0xffff00ffff00ffff, + 0x00ff00ffffff0000, 0xffff00ffffff00ff, 0xffff00ffffffff00, 0x00ff00ffffffffff, + 0x00ffff0000000000, 0xffffff00000000ff, 0xffffff000000ff00, 0x00ffff000000ffff, + 0xffffff0000ff0000, 0x00ffff0000ff00ff, 0x00ffff0000ffff00, 0xffffff0000ffffff, + 0xffffff00ff000000, 0x00ffff00ff0000ff, 0x00ffff00ff00ff00, 0xffffff00ff00ffff, + 0x00ffff00ffff0000, 0xffffff00ffff00ff, 0xffffff00ffffff00, 0x00ffff00ffffffff, + 0xffffffff00000000, 0x00ffffff000000ff, 0x00ffffff0000ff00, 0xffffffff0000ffff, + 0x00ffffff00ff0000, 0xffffffff00ff00ff, 0xffffffff00ffff00, 0x00ffffff00ffffff, + 0x00ffffffff000000, 0xffffffffff0000ff, 0xffffffffff00ff00, 0x00ffffffff00ffff, + 0xffffffffffff0000, 0x00ffffffffff00ff, 0x00ffffffffffff00, 0xffffffffffffffff, +GGML_TABLE_END() +//#endif + + +GGML_TABLE_BEGIN(uint64_t, iq2xxs_grid, 256) + 0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08, + 0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x08080808082b0808, + 0x08080808082b082b, 0x08080808082b2b08, 0x08080808082b2b2b, 0x0808080819080819, + 0x0808080819081908, 0x0808080819190808, 0x0808080819192b08, 0x08080808192b0819, + 0x08080808192b1908, 0x080808082b080808, 0x080808082b08082b, 0x080808082b082b2b, + 0x080808082b2b082b, 0x0808081908080819, 0x0808081908081908, 0x0808081908190808, + 0x0808081908191919, 0x0808081919080808, 0x080808192b081908, 0x080808192b192b08, + 0x0808082b08080808, 0x0808082b0808082b, 0x0808082b082b082b, 0x0808082b2b08082b, + 0x0808190808080819, 0x0808190808081908, 0x0808190808190808, 0x08081908082b0819, + 0x08081908082b1908, 0x0808190819080808, 0x080819081908082b, 0x0808190819082b08, + 0x08081908192b0808, 0x080819082b080819, 0x080819082b081908, 0x080819082b190808, + 0x080819082b2b1908, 0x0808191908080808, 0x080819190808082b, 0x0808191908082b08, + 0x08081919082b0808, 0x080819191908192b, 0x08081919192b2b19, 0x080819192b080808, + 0x080819192b190819, 0x0808192b08082b19, 0x0808192b08190808, 0x0808192b19080808, + 0x0808192b2b081908, 0x0808192b2b2b1908, 0x08082b0808080808, 0x08082b0808081919, + 0x08082b0808082b08, 0x08082b0808191908, 0x08082b08082b2b08, 0x08082b0819080819, + 0x08082b0819081908, 0x08082b0819190808, 0x08082b081919082b, 0x08082b082b082b08, + 0x08082b1908081908, 0x08082b1919080808, 0x08082b2b0808082b, 0x08082b2b08191908, + 0x0819080808080819, 0x0819080808081908, 0x0819080808190808, 0x08190808082b0819, + 0x0819080819080808, 0x08190808192b0808, 0x081908082b081908, 0x081908082b190808, + 0x081908082b191919, 0x0819081908080808, 0x0819081908082b08, 0x08190819082b0808, + 0x0819081919190808, 0x0819081919192b2b, 0x081908192b080808, 0x0819082b082b1908, + 0x0819082b19081919, 0x0819190808080808, 0x0819190808082b08, 0x08191908082b0808, + 0x08191908082b1919, 0x0819190819082b19, 0x081919082b080808, 0x0819191908192b08, + 0x08191919192b082b, 0x0819192b08080808, 0x0819192b0819192b, 0x08192b0808080819, + 0x08192b0808081908, 0x08192b0808190808, 0x08192b0819080808, 0x08192b082b080819, + 0x08192b1908080808, 0x08192b1908081919, 0x08192b192b2b0808, 0x08192b2b19190819, + 0x082b080808080808, 0x082b08080808082b, 0x082b080808082b2b, 0x082b080819081908, + 0x082b0808192b0819, 0x082b08082b080808, 0x082b08082b08082b, 0x082b0819082b2b19, + 0x082b081919082b08, 0x082b082b08080808, 0x082b082b0808082b, 0x082b190808080819, + 0x082b190808081908, 0x082b190808190808, 0x082b190819080808, 0x082b19081919192b, + 0x082b191908080808, 0x082b191919080819, 0x082b1919192b1908, 0x082b192b2b190808, + 0x082b2b0808082b08, 0x082b2b08082b0808, 0x082b2b082b191908, 0x082b2b2b19081908, + 0x1908080808080819, 0x1908080808081908, 0x1908080808190808, 0x1908080808192b08, + 0x19080808082b0819, 0x19080808082b1908, 0x1908080819080808, 0x1908080819082b08, + 0x190808081919192b, 0x19080808192b0808, 0x190808082b080819, 0x190808082b081908, + 0x190808082b190808, 0x1908081908080808, 0x19080819082b0808, 0x19080819192b0819, + 0x190808192b080808, 0x190808192b081919, 0x1908082b08080819, 0x1908082b08190808, + 0x1908082b19082b08, 0x1908082b1919192b, 0x1908082b192b2b08, 0x1908190808080808, + 0x1908190808082b08, 0x19081908082b0808, 0x190819082b080808, 0x190819082b192b19, + 0x190819190819082b, 0x19081919082b1908, 0x1908192b08080808, 0x19082b0808080819, + 0x19082b0808081908, 0x19082b0808190808, 0x19082b0819080808, 0x19082b0819081919, + 0x19082b1908080808, 0x19082b1919192b08, 0x19082b19192b0819, 0x19082b192b08082b, + 0x19082b2b19081919, 0x19082b2b2b190808, 0x1919080808080808, 0x1919080808082b08, + 0x1919080808190819, 0x1919080808192b19, 0x19190808082b0808, 0x191908082b080808, + 0x191908082b082b08, 0x1919081908081908, 0x191908191908082b, 0x191908192b2b1908, + 0x1919082b2b190819, 0x191919082b190808, 0x191919082b19082b, 0x1919191908082b2b, + 0x1919192b08080819, 0x1919192b19191908, 0x19192b0808080808, 0x19192b0808190819, + 0x19192b0808192b19, 0x19192b08192b1908, 0x19192b1919080808, 0x19192b2b08082b08, + 0x192b080808081908, 0x192b080808190808, 0x192b080819080808, 0x192b0808192b2b08, + 0x192b081908080808, 0x192b081919191919, 0x192b082b08192b08, 0x192b082b192b0808, + 0x192b190808080808, 0x192b190808081919, 0x192b191908190808, 0x192b19190819082b, + 0x192b19192b081908, 0x192b2b081908082b, 0x2b08080808080808, 0x2b0808080808082b, + 0x2b08080808082b2b, 0x2b08080819080819, 0x2b0808082b08082b, 0x2b08081908081908, + 0x2b08081908192b08, 0x2b08081919080808, 0x2b08082b08190819, 0x2b08190808080819, + 0x2b08190808081908, 0x2b08190808190808, 0x2b08190808191919, 0x2b08190819080808, + 0x2b081908192b0808, 0x2b08191908080808, 0x2b0819191908192b, 0x2b0819192b191908, + 0x2b08192b08082b19, 0x2b08192b19080808, 0x2b08192b192b0808, 0x2b082b080808082b, + 0x2b082b1908081908, 0x2b082b2b08190819, 0x2b19080808081908, 0x2b19080808190808, + 0x2b190808082b1908, 0x2b19080819080808, 0x2b1908082b2b0819, 0x2b1908190819192b, + 0x2b1908192b080808, 0x2b19082b19081919, 0x2b19190808080808, 0x2b191908082b082b, + 0x2b19190819081908, 0x2b19191919190819, 0x2b192b082b080819, 0x2b192b19082b0808, + 0x2b2b08080808082b, 0x2b2b080819190808, 0x2b2b08082b081919, 0x2b2b081908082b19, + 0x2b2b082b08080808, 0x2b2b190808192b08, 0x2b2b2b0819190808, 0x2b2b2b1908081908, +GGML_TABLE_END() + +GGML_TABLE_BEGIN(uint64_t, iq2xs_grid, 512) + 0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08, + 0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x080808080819192b, + 0x0808080808192b19, 0x08080808082b0808, 0x08080808082b082b, 0x08080808082b1919, + 0x08080808082b2b08, 0x0808080819080819, 0x0808080819081908, 0x080808081908192b, + 0x0808080819082b19, 0x0808080819190808, 0x080808081919082b, 0x0808080819191919, + 0x0808080819192b08, 0x08080808192b0819, 0x08080808192b1908, 0x080808082b080808, + 0x080808082b08082b, 0x080808082b081919, 0x080808082b082b08, 0x080808082b190819, + 0x080808082b191908, 0x080808082b192b19, 0x080808082b2b0808, 0x0808081908080819, + 0x0808081908081908, 0x080808190808192b, 0x0808081908082b19, 0x0808081908190808, + 0x080808190819082b, 0x0808081908191919, 0x0808081908192b08, 0x0808081908192b2b, + 0x08080819082b0819, 0x08080819082b1908, 0x0808081919080808, 0x080808191908082b, + 0x0808081919081919, 0x0808081919082b08, 0x0808081919190819, 0x0808081919191908, + 0x08080819192b0808, 0x08080819192b2b08, 0x080808192b080819, 0x080808192b081908, + 0x080808192b190808, 0x0808082b08080808, 0x0808082b0808082b, 0x0808082b08081919, + 0x0808082b08082b08, 0x0808082b08190819, 0x0808082b08191908, 0x0808082b082b0808, + 0x0808082b19080819, 0x0808082b19081908, 0x0808082b19190808, 0x0808082b19191919, + 0x0808082b2b080808, 0x0808082b2b082b2b, 0x0808190808080819, 0x0808190808081908, + 0x080819080808192b, 0x0808190808082b19, 0x0808190808190808, 0x080819080819082b, + 0x0808190808191919, 0x0808190808192b08, 0x08081908082b0819, 0x08081908082b1908, + 0x0808190819080808, 0x080819081908082b, 0x0808190819081919, 0x0808190819082b08, + 0x0808190819190819, 0x0808190819191908, 0x080819081919192b, 0x08081908192b0808, + 0x080819082b080819, 0x080819082b081908, 0x080819082b190808, 0x0808191908080808, + 0x080819190808082b, 0x0808191908081919, 0x0808191908082b08, 0x0808191908190819, + 0x0808191908191908, 0x08081919082b0808, 0x0808191919080819, 0x0808191919081908, + 0x0808191919190808, 0x08081919192b0819, 0x080819192b080808, 0x0808192b08080819, + 0x0808192b08081908, 0x0808192b08190808, 0x0808192b082b192b, 0x0808192b19080808, + 0x0808192b1908082b, 0x0808192b2b081908, 0x08082b0808080808, 0x08082b080808082b, + 0x08082b0808081919, 0x08082b0808082b08, 0x08082b0808082b2b, 0x08082b0808190819, + 0x08082b0808191908, 0x08082b08082b0808, 0x08082b08082b1919, 0x08082b0819080819, + 0x08082b0819081908, 0x08082b0819190808, 0x08082b0819192b08, 0x08082b082b080808, + 0x08082b082b2b0808, 0x08082b082b2b2b2b, 0x08082b1908080819, 0x08082b1908081908, + 0x08082b1908190808, 0x08082b1919080808, 0x08082b192b080819, 0x08082b192b082b19, + 0x08082b2b08080808, 0x08082b2b082b0808, 0x08082b2b082b2b08, 0x08082b2b2b19192b, + 0x08082b2b2b2b0808, 0x0819080808080819, 0x0819080808081908, 0x081908080808192b, + 0x0819080808082b19, 0x0819080808190808, 0x081908080819082b, 0x0819080808191919, + 0x0819080808192b08, 0x08190808082b0819, 0x08190808082b1908, 0x0819080819080808, + 0x081908081908082b, 0x0819080819081919, 0x0819080819082b08, 0x0819080819190819, + 0x0819080819191908, 0x08190808192b0808, 0x08190808192b2b2b, 0x081908082b080819, + 0x081908082b081908, 0x081908082b190808, 0x0819081908080808, 0x081908190808082b, + 0x0819081908081919, 0x0819081908082b08, 0x0819081908190819, 0x0819081908191908, + 0x08190819082b0808, 0x0819081919080819, 0x0819081919081908, 0x0819081919190808, + 0x081908192b080808, 0x081908192b191908, 0x081908192b19192b, 0x0819082b08080819, + 0x0819082b08081908, 0x0819082b0808192b, 0x0819082b08190808, 0x0819082b19080808, + 0x0819082b192b0808, 0x0819190808080808, 0x081919080808082b, 0x0819190808081919, + 0x0819190808082b08, 0x0819190808190819, 0x0819190808191908, 0x08191908082b0808, + 0x0819190819080819, 0x0819190819081908, 0x0819190819082b19, 0x0819190819190808, + 0x08191908192b1908, 0x081919082b080808, 0x0819191908080819, 0x0819191908081908, + 0x0819191908190808, 0x0819191919080808, 0x0819192b08080808, 0x0819192b08191908, + 0x0819192b19082b19, 0x08192b0808080819, 0x08192b0808081908, 0x08192b0808190808, + 0x08192b080819082b, 0x08192b0819080808, 0x08192b0819191908, 0x08192b082b08192b, + 0x08192b1908080808, 0x08192b1908081919, 0x08192b19192b192b, 0x08192b2b19190819, + 0x08192b2b2b2b2b19, 0x082b080808080808, 0x082b08080808082b, 0x082b080808081919, + 0x082b080808082b08, 0x082b080808082b2b, 0x082b080808190819, 0x082b080808191908, + 0x082b0808082b0808, 0x082b080819080819, 0x082b080819081908, 0x082b080819190808, + 0x082b08082b080808, 0x082b08082b2b0808, 0x082b081908080819, 0x082b081908081908, + 0x082b081908190808, 0x082b081919080808, 0x082b081919082b08, 0x082b0819192b1919, + 0x082b082b08080808, 0x082b082b082b082b, 0x082b082b2b080808, 0x082b082b2b2b2b08, + 0x082b190808080819, 0x082b190808081908, 0x082b190808190808, 0x082b1908082b2b19, + 0x082b190819080808, 0x082b191908080808, 0x082b191919080819, 0x082b19191919082b, + 0x082b19192b192b19, 0x082b192b08080819, 0x082b192b08192b2b, 0x082b192b2b2b192b, + 0x082b2b0808080808, 0x082b2b0808082b08, 0x082b2b0808082b2b, 0x082b2b08082b0808, + 0x082b2b0819191919, 0x082b2b082b082b08, 0x082b2b082b2b082b, 0x082b2b19192b2b08, + 0x082b2b192b190808, 0x082b2b2b08082b08, 0x082b2b2b082b0808, 0x082b2b2b2b08082b, + 0x082b2b2b2b082b08, 0x082b2b2b2b082b2b, 0x1908080808080819, 0x1908080808081908, + 0x190808080808192b, 0x1908080808082b19, 0x1908080808190808, 0x190808080819082b, + 0x1908080808191919, 0x1908080808192b08, 0x19080808082b0819, 0x19080808082b1908, + 0x1908080819080808, 0x190808081908082b, 0x1908080819081919, 0x1908080819082b08, + 0x1908080819082b2b, 0x1908080819190819, 0x1908080819191908, 0x19080808192b0808, + 0x19080808192b1919, 0x190808082b080819, 0x190808082b081908, 0x190808082b190808, + 0x1908081908080808, 0x190808190808082b, 0x1908081908081919, 0x1908081908082b08, + 0x1908081908190819, 0x1908081908191908, 0x19080819082b0808, 0x1908081919080819, + 0x1908081919081908, 0x1908081919190808, 0x190808192b080808, 0x190808192b081919, + 0x190808192b2b082b, 0x1908082b08080819, 0x1908082b08081908, 0x1908082b08190808, + 0x1908082b0819082b, 0x1908082b082b2b19, 0x1908082b19080808, 0x1908190808080808, + 0x190819080808082b, 0x1908190808081919, 0x1908190808082b08, 0x1908190808190819, + 0x1908190808191908, 0x1908190808192b19, 0x19081908082b0808, 0x1908190819080819, + 0x1908190819081908, 0x1908190819190808, 0x190819082b080808, 0x190819082b191908, + 0x1908191908080819, 0x1908191908081908, 0x1908191908190808, 0x19081919082b1908, + 0x1908191919080808, 0x190819192b192b2b, 0x1908192b08080808, 0x1908192b08082b2b, + 0x1908192b19081908, 0x1908192b19190808, 0x19082b0808080819, 0x19082b0808081908, + 0x19082b0808190808, 0x19082b0819080808, 0x19082b0819081919, 0x19082b0819191908, + 0x19082b08192b082b, 0x19082b1908080808, 0x19082b1908190819, 0x19082b1919081908, + 0x19082b1919190808, 0x19082b19192b2b19, 0x19082b2b08081908, 0x1919080808080808, + 0x191908080808082b, 0x1919080808081919, 0x1919080808082b08, 0x1919080808190819, + 0x1919080808191908, 0x19190808082b0808, 0x19190808082b2b08, 0x1919080819080819, + 0x1919080819081908, 0x1919080819190808, 0x191908082b080808, 0x1919081908080819, + 0x1919081908081908, 0x1919081908190808, 0x1919081908191919, 0x1919081919080808, + 0x191908191908082b, 0x1919082b08080808, 0x1919082b19081908, 0x1919082b2b2b2b2b, + 0x1919190808080819, 0x1919190808081908, 0x1919190808190808, 0x19191908082b0819, + 0x1919190819080808, 0x19191908192b0808, 0x191919082b080819, 0x191919082b2b0819, + 0x1919191908080808, 0x1919191908082b08, 0x191919192b080808, 0x191919192b082b08, + 0x1919192b082b0819, 0x1919192b192b2b08, 0x1919192b2b2b0819, 0x19192b0808080808, + 0x19192b0808191908, 0x19192b0819080819, 0x19192b0819190808, 0x19192b082b192b19, + 0x19192b1908192b2b, 0x19192b1919080808, 0x19192b191908082b, 0x19192b2b2b081919, + 0x192b080808080819, 0x192b080808081908, 0x192b080808190808, 0x192b080819080808, + 0x192b080819191908, 0x192b0808192b082b, 0x192b08082b08192b, 0x192b08082b2b2b19, + 0x192b081908080808, 0x192b082b082b1908, 0x192b082b19082b2b, 0x192b082b2b19082b, + 0x192b190808080808, 0x192b19080819192b, 0x192b191908190808, 0x192b191919080808, + 0x192b191919081919, 0x192b19192b2b1908, 0x192b2b0808080819, 0x192b2b08192b2b2b, + 0x192b2b19082b1919, 0x192b2b2b0808192b, 0x192b2b2b19191908, 0x192b2b2b192b082b, + 0x2b08080808080808, 0x2b0808080808082b, 0x2b08080808081919, 0x2b08080808082b08, + 0x2b08080808190819, 0x2b08080808191908, 0x2b080808082b0808, 0x2b080808082b2b2b, + 0x2b08080819080819, 0x2b08080819081908, 0x2b08080819190808, 0x2b0808082b080808, + 0x2b0808082b08082b, 0x2b0808082b2b2b08, 0x2b0808082b2b2b2b, 0x2b08081908080819, + 0x2b08081908081908, 0x2b0808190808192b, 0x2b08081908190808, 0x2b08081919080808, + 0x2b08081919190819, 0x2b08081919192b19, 0x2b08082b08080808, 0x2b08082b082b0808, + 0x2b08082b2b080808, 0x2b08082b2b08082b, 0x2b08082b2b2b0808, 0x2b08082b2b2b2b08, + 0x2b08190808080819, 0x2b08190808081908, 0x2b08190808190808, 0x2b0819080819082b, + 0x2b08190808191919, 0x2b08190819080808, 0x2b081908192b0808, 0x2b0819082b082b19, + 0x2b08191908080808, 0x2b08191919081908, 0x2b0819192b2b1919, 0x2b08192b08192b08, + 0x2b08192b192b2b2b, 0x2b082b0808080808, 0x2b082b0808082b08, 0x2b082b08082b1919, + 0x2b082b0819192b2b, 0x2b082b082b080808, 0x2b082b082b08082b, 0x2b082b082b2b2b08, + 0x2b082b190808192b, 0x2b082b2b082b082b, 0x2b082b2b2b080808, 0x2b082b2b2b082b08, + 0x2b082b2b2b19192b, 0x2b082b2b2b2b2b08, 0x2b19080808080819, 0x2b19080808081908, + 0x2b19080808190808, 0x2b19080819080808, 0x2b1908081919192b, 0x2b1908082b081908, + 0x2b19081908080808, 0x2b190819082b082b, 0x2b190819192b1908, 0x2b19082b1919192b, + 0x2b19082b2b082b19, 0x2b19190808080808, 0x2b19190808081919, 0x2b19190819081908, + 0x2b19190819190808, 0x2b19190819192b08, 0x2b191919082b2b19, 0x2b1919192b190808, + 0x2b1919192b19082b, 0x2b19192b19080819, 0x2b192b0819190819, 0x2b192b082b2b192b, + 0x2b192b1919082b19, 0x2b192b2b08191919, 0x2b192b2b192b0808, 0x2b2b080808080808, + 0x2b2b08080808082b, 0x2b2b080808082b08, 0x2b2b080808082b2b, 0x2b2b0808082b0808, + 0x2b2b0808082b2b2b, 0x2b2b08082b2b0808, 0x2b2b081919190819, 0x2b2b081919192b19, + 0x2b2b08192b2b192b, 0x2b2b082b08080808, 0x2b2b082b0808082b, 0x2b2b082b08082b08, + 0x2b2b082b082b2b2b, 0x2b2b082b2b080808, 0x2b2b082b2b2b0808, 0x2b2b190819080808, + 0x2b2b19082b191919, 0x2b2b192b192b1919, 0x2b2b192b2b192b08, 0x2b2b2b0808082b2b, + 0x2b2b2b08082b0808, 0x2b2b2b08082b082b, 0x2b2b2b08082b2b08, 0x2b2b2b082b2b0808, + 0x2b2b2b082b2b2b08, 0x2b2b2b1908081908, 0x2b2b2b192b081908, 0x2b2b2b192b08192b, + 0x2b2b2b2b082b2b08, 0x2b2b2b2b082b2b2b, 0x2b2b2b2b2b190819, 0x2b2b2b2b2b2b2b2b, +GGML_TABLE_END() + +GGML_TABLE_BEGIN(uint64_t, iq2s_grid, 1024) + 0x0808080808080808, 0x080808080808082b, 0x0808080808081919, 0x0808080808082b08, + 0x0808080808082b2b, 0x0808080808190819, 0x0808080808191908, 0x080808080819192b, + 0x0808080808192b19, 0x08080808082b0808, 0x08080808082b082b, 0x08080808082b1919, + 0x08080808082b2b08, 0x0808080819080819, 0x0808080819081908, 0x080808081908192b, + 0x0808080819082b19, 0x0808080819190808, 0x080808081919082b, 0x0808080819191919, + 0x0808080819192b08, 0x08080808192b0819, 0x08080808192b1908, 0x08080808192b192b, + 0x08080808192b2b19, 0x080808082b080808, 0x080808082b08082b, 0x080808082b081919, + 0x080808082b082b08, 0x080808082b190819, 0x080808082b191908, 0x080808082b2b0808, + 0x080808082b2b1919, 0x080808082b2b2b2b, 0x0808081908080819, 0x0808081908081908, + 0x080808190808192b, 0x0808081908082b19, 0x0808081908190808, 0x080808190819082b, + 0x0808081908191919, 0x0808081908192b08, 0x08080819082b0819, 0x08080819082b1908, + 0x0808081919080808, 0x080808191908082b, 0x0808081919081919, 0x0808081919082b08, + 0x0808081919190819, 0x0808081919191908, 0x080808191919192b, 0x0808081919192b19, + 0x08080819192b0808, 0x08080819192b1919, 0x08080819192b2b08, 0x080808192b080819, + 0x080808192b081908, 0x080808192b190808, 0x080808192b19082b, 0x080808192b191919, + 0x080808192b2b0819, 0x080808192b2b1908, 0x0808082b08080808, 0x0808082b0808082b, + 0x0808082b08081919, 0x0808082b08082b08, 0x0808082b08190819, 0x0808082b08191908, + 0x0808082b082b0808, 0x0808082b082b2b2b, 0x0808082b19080819, 0x0808082b19081908, + 0x0808082b1908192b, 0x0808082b19082b19, 0x0808082b19190808, 0x0808082b19191919, + 0x0808082b2b080808, 0x0808082b2b081919, 0x0808082b2b082b2b, 0x0808082b2b191908, + 0x0808082b2b2b082b, 0x0808190808080819, 0x0808190808081908, 0x080819080808192b, + 0x0808190808082b19, 0x0808190808190808, 0x080819080819082b, 0x0808190808191919, + 0x0808190808192b08, 0x08081908082b0819, 0x08081908082b1908, 0x08081908082b192b, + 0x08081908082b2b19, 0x0808190819080808, 0x080819081908082b, 0x0808190819081919, + 0x0808190819082b08, 0x0808190819082b2b, 0x0808190819190819, 0x0808190819191908, + 0x080819081919192b, 0x0808190819192b19, 0x08081908192b0808, 0x08081908192b082b, + 0x08081908192b1919, 0x080819082b080819, 0x080819082b081908, 0x080819082b08192b, + 0x080819082b082b19, 0x080819082b190808, 0x080819082b191919, 0x080819082b192b08, + 0x080819082b2b0819, 0x080819082b2b1908, 0x0808191908080808, 0x080819190808082b, + 0x0808191908081919, 0x0808191908082b08, 0x0808191908082b2b, 0x0808191908190819, + 0x0808191908191908, 0x080819190819192b, 0x0808191908192b19, 0x08081919082b0808, + 0x08081919082b1919, 0x08081919082b2b08, 0x0808191919080819, 0x0808191919081908, + 0x080819191908192b, 0x0808191919082b19, 0x0808191919190808, 0x080819191919082b, + 0x0808191919191919, 0x0808191919192b08, 0x08081919192b0819, 0x08081919192b1908, + 0x080819192b080808, 0x080819192b08082b, 0x080819192b081919, 0x080819192b082b08, + 0x080819192b190819, 0x080819192b191908, 0x080819192b2b0808, 0x0808192b08080819, + 0x0808192b08081908, 0x0808192b0808192b, 0x0808192b08082b19, 0x0808192b08190808, + 0x0808192b08191919, 0x0808192b19080808, 0x0808192b19081919, 0x0808192b19082b08, + 0x0808192b19190819, 0x0808192b19191908, 0x0808192b192b0808, 0x0808192b2b080819, + 0x0808192b2b081908, 0x0808192b2b190808, 0x08082b0808080808, 0x08082b080808082b, + 0x08082b0808081919, 0x08082b0808082b08, 0x08082b0808190819, 0x08082b0808191908, + 0x08082b080819192b, 0x08082b0808192b19, 0x08082b08082b0808, 0x08082b08082b1919, + 0x08082b08082b2b2b, 0x08082b0819080819, 0x08082b0819081908, 0x08082b081908192b, + 0x08082b0819082b19, 0x08082b0819190808, 0x08082b081919082b, 0x08082b0819191919, + 0x08082b0819192b08, 0x08082b08192b0819, 0x08082b08192b1908, 0x08082b082b080808, + 0x08082b082b081919, 0x08082b082b191908, 0x08082b082b2b2b2b, 0x08082b1908080819, + 0x08082b1908081908, 0x08082b1908190808, 0x08082b190819082b, 0x08082b1908191919, + 0x08082b1908192b08, 0x08082b19082b0819, 0x08082b1919080808, 0x08082b1919081919, + 0x08082b1919082b08, 0x08082b1919190819, 0x08082b1919191908, 0x08082b19192b0808, + 0x08082b192b080819, 0x08082b192b190808, 0x08082b2b08080808, 0x08082b2b08190819, + 0x08082b2b08191908, 0x08082b2b082b082b, 0x08082b2b082b2b08, 0x08082b2b082b2b2b, + 0x08082b2b19190808, 0x08082b2b2b192b19, 0x0819080808080819, 0x0819080808081908, + 0x081908080808192b, 0x0819080808082b19, 0x0819080808190808, 0x081908080819082b, + 0x0819080808191919, 0x0819080808192b08, 0x08190808082b0819, 0x08190808082b1908, + 0x08190808082b192b, 0x0819080819080808, 0x081908081908082b, 0x0819080819081919, + 0x0819080819082b08, 0x0819080819190819, 0x0819080819191908, 0x081908081919192b, + 0x0819080819192b19, 0x08190808192b0808, 0x08190808192b082b, 0x08190808192b1919, + 0x08190808192b2b08, 0x081908082b080819, 0x081908082b081908, 0x081908082b08192b, + 0x081908082b190808, 0x081908082b191919, 0x081908082b192b08, 0x081908082b2b0819, + 0x081908082b2b1908, 0x0819081908080808, 0x081908190808082b, 0x0819081908081919, + 0x0819081908082b08, 0x0819081908082b2b, 0x0819081908190819, 0x0819081908191908, + 0x081908190819192b, 0x0819081908192b19, 0x08190819082b0808, 0x08190819082b082b, + 0x08190819082b1919, 0x08190819082b2b08, 0x0819081919080819, 0x0819081919081908, + 0x081908191908192b, 0x0819081919082b19, 0x0819081919190808, 0x081908191919082b, + 0x0819081919191919, 0x0819081919192b08, 0x08190819192b0819, 0x08190819192b1908, + 0x081908192b080808, 0x081908192b08082b, 0x081908192b081919, 0x081908192b082b08, + 0x081908192b190819, 0x081908192b191908, 0x0819082b08080819, 0x0819082b08081908, + 0x0819082b08082b19, 0x0819082b08190808, 0x0819082b08191919, 0x0819082b082b0819, + 0x0819082b082b1908, 0x0819082b19080808, 0x0819082b19081919, 0x0819082b19190819, + 0x0819082b19191908, 0x0819082b2b080819, 0x0819082b2b081908, 0x0819082b2b190808, + 0x0819190808080808, 0x081919080808082b, 0x0819190808081919, 0x0819190808082b08, + 0x0819190808190819, 0x0819190808191908, 0x081919080819192b, 0x0819190808192b19, + 0x08191908082b0808, 0x08191908082b1919, 0x08191908082b2b08, 0x0819190819080819, + 0x0819190819081908, 0x081919081908192b, 0x0819190819082b19, 0x0819190819190808, + 0x081919081919082b, 0x0819190819191919, 0x0819190819192b08, 0x08191908192b0819, + 0x08191908192b1908, 0x081919082b080808, 0x081919082b08082b, 0x081919082b081919, + 0x081919082b082b08, 0x081919082b190819, 0x081919082b191908, 0x081919082b2b0808, + 0x0819191908080819, 0x0819191908081908, 0x081919190808192b, 0x0819191908082b19, + 0x0819191908190808, 0x081919190819082b, 0x0819191908191919, 0x0819191908192b08, + 0x08191919082b0819, 0x08191919082b1908, 0x0819191919080808, 0x081919191908082b, + 0x0819191919081919, 0x0819191919082b08, 0x0819191919190819, 0x0819191919191908, + 0x08191919192b0808, 0x081919192b080819, 0x081919192b081908, 0x081919192b190808, + 0x0819192b08080808, 0x0819192b08081919, 0x0819192b08082b08, 0x0819192b08190819, + 0x0819192b08191908, 0x0819192b082b0808, 0x0819192b19080819, 0x0819192b19081908, + 0x0819192b19190808, 0x0819192b2b080808, 0x0819192b2b2b2b2b, 0x08192b0808080819, + 0x08192b0808081908, 0x08192b080808192b, 0x08192b0808082b19, 0x08192b0808190808, + 0x08192b0808191919, 0x08192b0808192b08, 0x08192b08082b0819, 0x08192b0819080808, + 0x08192b081908082b, 0x08192b0819081919, 0x08192b0819082b08, 0x08192b0819190819, + 0x08192b0819191908, 0x08192b08192b0808, 0x08192b082b080819, 0x08192b082b081908, + 0x08192b1908080808, 0x08192b190808082b, 0x08192b1908081919, 0x08192b1908082b08, + 0x08192b1908190819, 0x08192b1908191908, 0x08192b19082b0808, 0x08192b1919080819, + 0x08192b1919081908, 0x08192b1919190808, 0x08192b19192b2b19, 0x08192b192b2b082b, + 0x08192b2b08081908, 0x08192b2b08190808, 0x08192b2b19080808, 0x08192b2b1919192b, + 0x082b080808080808, 0x082b08080808082b, 0x082b080808081919, 0x082b080808082b08, + 0x082b080808190819, 0x082b080808191908, 0x082b08080819192b, 0x082b080808192b19, + 0x082b0808082b0808, 0x082b0808082b1919, 0x082b0808082b2b2b, 0x082b080819080819, + 0x082b080819081908, 0x082b080819190808, 0x082b08081919082b, 0x082b080819191919, + 0x082b0808192b1908, 0x082b08082b080808, 0x082b08082b082b2b, 0x082b08082b191908, + 0x082b08082b2b2b2b, 0x082b081908080819, 0x082b081908081908, 0x082b081908190808, + 0x082b08190819082b, 0x082b081908191919, 0x082b0819082b0819, 0x082b081919080808, + 0x082b08191908082b, 0x082b081919081919, 0x082b081919190819, 0x082b081919191908, + 0x082b0819192b0808, 0x082b08192b080819, 0x082b08192b081908, 0x082b08192b190808, + 0x082b082b08080808, 0x082b082b08082b2b, 0x082b082b082b082b, 0x082b082b082b2b08, + 0x082b082b082b2b2b, 0x082b082b19081908, 0x082b082b19190808, 0x082b082b2b082b08, + 0x082b082b2b082b2b, 0x082b082b2b2b2b08, 0x082b190808080819, 0x082b190808081908, + 0x082b19080808192b, 0x082b190808082b19, 0x082b190808190808, 0x082b190808191919, + 0x082b190808192b08, 0x082b1908082b0819, 0x082b1908082b1908, 0x082b190819080808, + 0x082b19081908082b, 0x082b190819081919, 0x082b190819082b08, 0x082b190819190819, + 0x082b190819191908, 0x082b1908192b0808, 0x082b19082b080819, 0x082b19082b081908, + 0x082b19082b190808, 0x082b191908080808, 0x082b191908081919, 0x082b191908082b08, + 0x082b191908190819, 0x082b191908191908, 0x082b1919082b0808, 0x082b191919080819, + 0x082b191919081908, 0x082b191919190808, 0x082b1919192b192b, 0x082b19192b080808, + 0x082b192b08080819, 0x082b192b08081908, 0x082b192b08190808, 0x082b192b19080808, + 0x082b192b19192b19, 0x082b2b0808080808, 0x082b2b0808081919, 0x082b2b0808190819, + 0x082b2b0808191908, 0x082b2b0819080819, 0x082b2b0819081908, 0x082b2b0819190808, + 0x082b2b082b082b2b, 0x082b2b082b2b2b2b, 0x082b2b1908080819, 0x082b2b1908081908, + 0x082b2b1908190808, 0x082b2b192b191919, 0x082b2b2b08082b2b, 0x082b2b2b082b082b, + 0x082b2b2b192b1908, 0x082b2b2b2b082b08, 0x082b2b2b2b082b2b, 0x1908080808080819, + 0x1908080808081908, 0x190808080808192b, 0x1908080808082b19, 0x1908080808190808, + 0x190808080819082b, 0x1908080808191919, 0x1908080808192b08, 0x1908080808192b2b, + 0x19080808082b0819, 0x19080808082b1908, 0x19080808082b192b, 0x1908080819080808, + 0x190808081908082b, 0x1908080819081919, 0x1908080819082b08, 0x1908080819082b2b, + 0x1908080819190819, 0x1908080819191908, 0x190808081919192b, 0x1908080819192b19, + 0x19080808192b0808, 0x19080808192b082b, 0x19080808192b1919, 0x190808082b080819, + 0x190808082b081908, 0x190808082b190808, 0x190808082b191919, 0x190808082b192b08, + 0x190808082b2b0819, 0x190808082b2b1908, 0x1908081908080808, 0x190808190808082b, + 0x1908081908081919, 0x1908081908082b08, 0x1908081908190819, 0x1908081908191908, + 0x190808190819192b, 0x1908081908192b19, 0x19080819082b0808, 0x19080819082b082b, + 0x19080819082b1919, 0x1908081919080819, 0x1908081919081908, 0x190808191908192b, + 0x1908081919082b19, 0x1908081919190808, 0x190808191919082b, 0x1908081919191919, + 0x1908081919192b08, 0x19080819192b0819, 0x19080819192b1908, 0x190808192b080808, + 0x190808192b08082b, 0x190808192b081919, 0x190808192b082b08, 0x190808192b190819, + 0x190808192b191908, 0x190808192b2b0808, 0x1908082b08080819, 0x1908082b08081908, + 0x1908082b08190808, 0x1908082b0819082b, 0x1908082b08191919, 0x1908082b08192b08, + 0x1908082b082b1908, 0x1908082b19080808, 0x1908082b19081919, 0x1908082b19082b08, + 0x1908082b19190819, 0x1908082b19191908, 0x1908082b192b0808, 0x1908082b2b080819, + 0x1908082b2b081908, 0x1908190808080808, 0x190819080808082b, 0x1908190808081919, + 0x1908190808082b08, 0x1908190808082b2b, 0x1908190808190819, 0x1908190808191908, + 0x190819080819192b, 0x1908190808192b19, 0x19081908082b0808, 0x19081908082b082b, + 0x19081908082b1919, 0x19081908082b2b08, 0x1908190819080819, 0x1908190819081908, + 0x190819081908192b, 0x1908190819082b19, 0x1908190819190808, 0x190819081919082b, + 0x1908190819191919, 0x1908190819192b08, 0x19081908192b0819, 0x19081908192b1908, + 0x190819082b080808, 0x190819082b08082b, 0x190819082b081919, 0x190819082b082b08, + 0x190819082b190819, 0x190819082b191908, 0x190819082b2b0808, 0x1908191908080819, + 0x1908191908081908, 0x190819190808192b, 0x1908191908082b19, 0x1908191908190808, + 0x190819190819082b, 0x1908191908191919, 0x1908191908192b08, 0x19081919082b0819, + 0x19081919082b1908, 0x1908191919080808, 0x190819191908082b, 0x1908191919081919, + 0x1908191919082b08, 0x1908191919190819, 0x1908191919191908, 0x19081919192b0808, + 0x19081919192b2b2b, 0x190819192b080819, 0x190819192b081908, 0x190819192b190808, + 0x1908192b08080808, 0x1908192b0808082b, 0x1908192b08081919, 0x1908192b08082b08, + 0x1908192b08190819, 0x1908192b08191908, 0x1908192b082b0808, 0x1908192b19080819, + 0x1908192b19081908, 0x1908192b19190808, 0x1908192b2b080808, 0x1908192b2b2b1919, + 0x19082b0808080819, 0x19082b0808081908, 0x19082b0808082b19, 0x19082b0808190808, + 0x19082b080819082b, 0x19082b0808191919, 0x19082b0808192b08, 0x19082b08082b0819, + 0x19082b08082b1908, 0x19082b0819080808, 0x19082b081908082b, 0x19082b0819081919, + 0x19082b0819082b08, 0x19082b0819190819, 0x19082b0819191908, 0x19082b08192b0808, + 0x19082b082b081908, 0x19082b082b190808, 0x19082b1908080808, 0x19082b190808082b, + 0x19082b1908081919, 0x19082b1908082b08, 0x19082b1908190819, 0x19082b1908191908, + 0x19082b19082b0808, 0x19082b1919080819, 0x19082b1919081908, 0x19082b1919190808, + 0x19082b192b080808, 0x19082b192b19192b, 0x19082b2b08080819, 0x19082b2b08081908, + 0x19082b2b08190808, 0x19082b2b19080808, 0x1919080808080808, 0x191908080808082b, + 0x1919080808081919, 0x1919080808082b08, 0x1919080808190819, 0x1919080808191908, + 0x191908080819192b, 0x1919080808192b19, 0x19190808082b0808, 0x19190808082b082b, + 0x19190808082b1919, 0x19190808082b2b08, 0x1919080819080819, 0x1919080819081908, + 0x191908081908192b, 0x1919080819082b19, 0x1919080819190808, 0x191908081919082b, + 0x1919080819191919, 0x1919080819192b08, 0x19190808192b0819, 0x19190808192b1908, + 0x191908082b080808, 0x191908082b08082b, 0x191908082b081919, 0x191908082b082b08, + 0x191908082b190819, 0x191908082b191908, 0x1919081908080819, 0x1919081908081908, + 0x191908190808192b, 0x1919081908082b19, 0x1919081908190808, 0x191908190819082b, + 0x1919081908191919, 0x1919081908192b08, 0x19190819082b0819, 0x19190819082b1908, + 0x1919081919080808, 0x191908191908082b, 0x1919081919081919, 0x1919081919082b08, + 0x1919081919190819, 0x1919081919191908, 0x19190819192b0808, 0x191908192b080819, + 0x191908192b081908, 0x191908192b190808, 0x1919082b08080808, 0x1919082b08081919, + 0x1919082b08082b08, 0x1919082b08190819, 0x1919082b08191908, 0x1919082b082b0808, + 0x1919082b19080819, 0x1919082b19081908, 0x1919082b19190808, 0x1919082b192b2b19, + 0x1919082b2b080808, 0x1919190808080819, 0x1919190808081908, 0x191919080808192b, + 0x1919190808082b19, 0x1919190808190808, 0x191919080819082b, 0x1919190808191919, + 0x1919190808192b08, 0x19191908082b0819, 0x19191908082b1908, 0x1919190819080808, + 0x191919081908082b, 0x1919190819081919, 0x1919190819082b08, 0x1919190819190819, + 0x1919190819191908, 0x19191908192b0808, 0x191919082b080819, 0x191919082b081908, + 0x191919082b190808, 0x1919191908080808, 0x191919190808082b, 0x1919191908081919, + 0x1919191908082b08, 0x1919191908190819, 0x1919191908191908, 0x19191919082b0808, + 0x1919191919080819, 0x1919191919081908, 0x1919191919190808, 0x191919192b080808, + 0x1919192b08080819, 0x1919192b08081908, 0x1919192b08190808, 0x1919192b082b192b, + 0x1919192b19080808, 0x19192b0808080808, 0x19192b080808082b, 0x19192b0808081919, + 0x19192b0808082b08, 0x19192b0808190819, 0x19192b0808191908, 0x19192b08082b0808, + 0x19192b0819080819, 0x19192b0819081908, 0x19192b0819190808, 0x19192b0819192b2b, + 0x19192b082b080808, 0x19192b1908080819, 0x19192b1908081908, 0x19192b1908190808, + 0x19192b1919080808, 0x19192b2b08080808, 0x19192b2b08192b19, 0x19192b2b2b081919, + 0x19192b2b2b2b2b08, 0x192b080808080819, 0x192b080808081908, 0x192b08080808192b, + 0x192b080808190808, 0x192b08080819082b, 0x192b080808191919, 0x192b080808192b08, + 0x192b0808082b0819, 0x192b0808082b1908, 0x192b080819080808, 0x192b080819081919, + 0x192b080819082b08, 0x192b080819190819, 0x192b080819191908, 0x192b0808192b0808, + 0x192b08082b081908, 0x192b08082b190808, 0x192b081908080808, 0x192b08190808082b, + 0x192b081908081919, 0x192b081908082b08, 0x192b081908190819, 0x192b081908191908, + 0x192b0819082b0808, 0x192b081919080819, 0x192b081919081908, 0x192b081919190808, + 0x192b08192b080808, 0x192b08192b192b19, 0x192b082b08081908, 0x192b082b08190808, + 0x192b082b19080808, 0x192b082b1919192b, 0x192b082b2b2b0819, 0x192b190808080808, + 0x192b190808081919, 0x192b190808082b08, 0x192b190808190819, 0x192b190808191908, + 0x192b1908082b0808, 0x192b190819080819, 0x192b190819081908, 0x192b190819190808, + 0x192b19082b080808, 0x192b191908080819, 0x192b191908081908, 0x192b191908190808, + 0x192b191919080808, 0x192b191919082b2b, 0x192b1919192b2b08, 0x192b19192b19082b, + 0x192b192b08080808, 0x192b192b2b191908, 0x192b2b0808080819, 0x192b2b0808081908, + 0x192b2b0808190808, 0x192b2b08192b1919, 0x192b2b082b192b08, 0x192b2b1908080808, + 0x192b2b19082b2b2b, 0x192b2b2b1908082b, 0x192b2b2b2b2b0819, 0x2b08080808080808, + 0x2b0808080808082b, 0x2b08080808081919, 0x2b08080808082b08, 0x2b08080808190819, + 0x2b08080808191908, 0x2b08080808192b19, 0x2b080808082b0808, 0x2b080808082b1919, + 0x2b08080819080819, 0x2b08080819081908, 0x2b08080819190808, 0x2b0808081919082b, + 0x2b08080819191919, 0x2b08080819192b08, 0x2b080808192b0819, 0x2b0808082b080808, + 0x2b0808082b081919, 0x2b0808082b190819, 0x2b0808082b191908, 0x2b08081908080819, + 0x2b08081908081908, 0x2b08081908082b19, 0x2b08081908190808, 0x2b0808190819082b, + 0x2b08081908191919, 0x2b08081908192b08, 0x2b080819082b0819, 0x2b080819082b1908, + 0x2b08081919080808, 0x2b0808191908082b, 0x2b08081919081919, 0x2b08081919082b08, + 0x2b08081919190819, 0x2b08081919191908, 0x2b0808192b080819, 0x2b0808192b081908, + 0x2b0808192b190808, 0x2b0808192b2b2b19, 0x2b08082b08080808, 0x2b08082b08081919, + 0x2b08082b08082b2b, 0x2b08082b08190819, 0x2b08082b08191908, 0x2b08082b19080819, + 0x2b08082b19081908, 0x2b08082b19190808, 0x2b08190808080819, 0x2b08190808081908, + 0x2b0819080808192b, 0x2b08190808082b19, 0x2b08190808190808, 0x2b0819080819082b, + 0x2b08190808191919, 0x2b08190808192b08, 0x2b081908082b0819, 0x2b08190819080808, + 0x2b0819081908082b, 0x2b08190819081919, 0x2b08190819082b08, 0x2b08190819190819, + 0x2b08190819191908, 0x2b081908192b0808, 0x2b0819082b080819, 0x2b0819082b081908, + 0x2b0819082b190808, 0x2b08191908080808, 0x2b0819190808082b, 0x2b08191908081919, + 0x2b08191908082b08, 0x2b08191908190819, 0x2b08191908191908, 0x2b081919082b0808, + 0x2b08191919080819, 0x2b08191919081908, 0x2b08191919190808, 0x2b0819192b080808, + 0x2b0819192b082b2b, 0x2b08192b08080819, 0x2b08192b08081908, 0x2b08192b08190808, + 0x2b08192b082b2b19, 0x2b08192b19080808, 0x2b082b0808080808, 0x2b082b0808081919, + 0x2b082b0808190819, 0x2b082b0808191908, 0x2b082b0819080819, 0x2b082b0819081908, + 0x2b082b0819190808, 0x2b082b082b2b082b, 0x2b082b1908080819, 0x2b082b1908081908, + 0x2b082b1919080808, 0x2b082b19192b1919, 0x2b082b2b082b082b, 0x2b082b2b19192b08, + 0x2b082b2b19192b2b, 0x2b082b2b2b08082b, 0x2b082b2b2b2b082b, 0x2b19080808080819, + 0x2b19080808081908, 0x2b19080808082b19, 0x2b19080808190808, 0x2b1908080819082b, + 0x2b19080808191919, 0x2b19080808192b08, 0x2b190808082b1908, 0x2b19080819080808, + 0x2b1908081908082b, 0x2b19080819081919, 0x2b19080819082b08, 0x2b19080819190819, + 0x2b19080819191908, 0x2b190808192b0808, 0x2b1908082b080819, 0x2b1908082b081908, + 0x2b1908082b190808, 0x2b19081908080808, 0x2b19081908081919, 0x2b19081908190819, + 0x2b19081908191908, 0x2b19081919080819, 0x2b19081919081908, 0x2b19081919190808, + 0x2b19081919192b2b, 0x2b19082b08080819, 0x2b19082b08081908, 0x2b19082b08190808, + 0x2b19082b19080808, 0x2b19082b2b2b192b, 0x2b19190808080808, 0x2b1919080808082b, + 0x2b19190808081919, 0x2b19190808082b08, 0x2b19190808190819, 0x2b19190808191908, + 0x2b191908082b0808, 0x2b19190819080819, 0x2b19190819081908, 0x2b19190819190808, + 0x2b1919082b080808, 0x2b1919082b19192b, 0x2b19191908080819, 0x2b19191908081908, + 0x2b19191908190808, 0x2b19191919080808, 0x2b1919192b192b08, 0x2b1919192b2b0819, + 0x2b19192b08080808, 0x2b19192b1908192b, 0x2b19192b192b1908, 0x2b192b0808080819, + 0x2b192b0808081908, 0x2b192b0808190808, 0x2b192b08082b192b, 0x2b192b0819080808, + 0x2b192b082b2b2b19, 0x2b192b1908080808, 0x2b192b1919082b19, 0x2b192b191919082b, + 0x2b192b2b2b190808, 0x2b2b080808080808, 0x2b2b080808081919, 0x2b2b080808082b2b, + 0x2b2b080808191908, 0x2b2b0808082b082b, 0x2b2b0808082b2b2b, 0x2b2b080819080819, + 0x2b2b080819081908, 0x2b2b080819190808, 0x2b2b08082b2b082b, 0x2b2b08082b2b2b2b, + 0x2b2b081919080808, 0x2b2b0819192b1919, 0x2b2b082b0808082b, 0x2b2b082b08082b2b, + 0x2b2b082b082b082b, 0x2b2b082b082b2b08, 0x2b2b082b082b2b2b, 0x2b2b082b2b08082b, + 0x2b2b082b2b082b08, 0x2b2b082b2b082b2b, 0x2b2b082b2b2b2b08, 0x2b2b190808080819, + 0x2b2b190808081908, 0x2b2b190808190808, 0x2b2b190819080808, 0x2b2b19082b082b19, + 0x2b2b19082b2b1908, 0x2b2b191908080808, 0x2b2b191908192b19, 0x2b2b192b19190819, + 0x2b2b2b0808082b2b, 0x2b2b2b08082b2b08, 0x2b2b2b082b2b082b, 0x2b2b2b1919191908, + 0x2b2b2b192b08192b, 0x2b2b2b2b08082b08, 0x2b2b2b2b08082b2b, 0x2b2b2b2b082b0808, + 0x2b2b2b2b082b082b, 0x2b2b2b2b082b2b08, 0x2b2b2b2b2b082b08, 0x2b2b2b2b2b2b2b2b, +GGML_TABLE_END() + +GGML_TABLE_BEGIN(uint32_t, iq3xxs_grid, 256) + 0x04040404, 0x04040414, 0x04040424, 0x04040c0c, 0x04040c1c, 0x04040c3e, 0x04041404, 0x04041414, + 0x04041c0c, 0x04042414, 0x04043e1c, 0x04043e2c, 0x040c040c, 0x040c041c, 0x040c0c04, 0x040c0c14, + 0x040c140c, 0x040c142c, 0x040c1c04, 0x040c1c14, 0x040c240c, 0x040c2c24, 0x040c3e04, 0x04140404, + 0x04140414, 0x04140424, 0x04140c0c, 0x04141404, 0x04141414, 0x04141c0c, 0x04141c1c, 0x04141c3e, + 0x04142c0c, 0x04142c3e, 0x04143e2c, 0x041c040c, 0x041c043e, 0x041c0c04, 0x041c0c14, 0x041c142c, + 0x041c3e04, 0x04240c1c, 0x04241c3e, 0x04242424, 0x04242c3e, 0x04243e1c, 0x04243e2c, 0x042c040c, + 0x042c043e, 0x042c1c14, 0x042c2c14, 0x04341c2c, 0x04343424, 0x043e0c04, 0x043e0c24, 0x043e0c34, + 0x043e241c, 0x043e340c, 0x0c04040c, 0x0c04041c, 0x0c040c04, 0x0c040c14, 0x0c04140c, 0x0c04141c, + 0x0c041c04, 0x0c041c14, 0x0c041c24, 0x0c04243e, 0x0c042c04, 0x0c0c0404, 0x0c0c0414, 0x0c0c0c0c, + 0x0c0c1404, 0x0c0c1414, 0x0c14040c, 0x0c14041c, 0x0c140c04, 0x0c140c14, 0x0c14140c, 0x0c141c04, + 0x0c143e14, 0x0c1c0404, 0x0c1c0414, 0x0c1c1404, 0x0c1c1c0c, 0x0c1c2434, 0x0c1c3434, 0x0c24040c, + 0x0c24042c, 0x0c242c04, 0x0c2c1404, 0x0c2c1424, 0x0c2c2434, 0x0c2c3e0c, 0x0c34042c, 0x0c3e1414, + 0x0c3e2404, 0x14040404, 0x14040414, 0x14040c0c, 0x14040c1c, 0x14041404, 0x14041414, 0x14041434, + 0x14041c0c, 0x14042414, 0x140c040c, 0x140c041c, 0x140c042c, 0x140c0c04, 0x140c0c14, 0x140c140c, + 0x140c1c04, 0x140c341c, 0x140c343e, 0x140c3e04, 0x14140404, 0x14140414, 0x14140c0c, 0x14140c3e, + 0x14141404, 0x14141414, 0x14141c3e, 0x14142404, 0x14142c2c, 0x141c040c, 0x141c0c04, 0x141c0c24, + 0x141c3e04, 0x141c3e24, 0x14241c2c, 0x14242c1c, 0x142c041c, 0x142c143e, 0x142c240c, 0x142c3e24, + 0x143e040c, 0x143e041c, 0x143e0c34, 0x143e242c, 0x1c04040c, 0x1c040c04, 0x1c040c14, 0x1c04140c, + 0x1c04141c, 0x1c042c04, 0x1c04342c, 0x1c043e14, 0x1c0c0404, 0x1c0c0414, 0x1c0c1404, 0x1c0c1c0c, + 0x1c0c2424, 0x1c0c2434, 0x1c14040c, 0x1c14041c, 0x1c140c04, 0x1c14142c, 0x1c142c14, 0x1c143e14, + 0x1c1c0c0c, 0x1c1c1c1c, 0x1c241c04, 0x1c24243e, 0x1c243e14, 0x1c2c0404, 0x1c2c0434, 0x1c2c1414, + 0x1c2c2c2c, 0x1c340c24, 0x1c341c34, 0x1c34341c, 0x1c3e1c1c, 0x1c3e3404, 0x24040424, 0x24040c3e, + 0x24041c2c, 0x24041c3e, 0x24042c1c, 0x24042c3e, 0x240c3e24, 0x24141404, 0x24141c3e, 0x24142404, + 0x24143404, 0x24143434, 0x241c043e, 0x241c242c, 0x24240424, 0x24242c0c, 0x24243424, 0x242c142c, + 0x242c241c, 0x242c3e04, 0x243e042c, 0x243e0c04, 0x243e0c14, 0x243e1c04, 0x2c040c14, 0x2c04240c, + 0x2c043e04, 0x2c0c0404, 0x2c0c0434, 0x2c0c1434, 0x2c0c2c2c, 0x2c140c24, 0x2c141c14, 0x2c143e14, + 0x2c1c0414, 0x2c1c2c1c, 0x2c240c04, 0x2c24141c, 0x2c24143e, 0x2c243e14, 0x2c2c0414, 0x2c2c1c0c, + 0x2c342c04, 0x2c3e1424, 0x2c3e2414, 0x34041424, 0x34042424, 0x34042434, 0x34043424, 0x340c140c, + 0x340c340c, 0x34140c3e, 0x34143424, 0x341c1c04, 0x341c1c34, 0x34242424, 0x342c042c, 0x342c2c14, + 0x34341c1c, 0x343e041c, 0x343e140c, 0x3e04041c, 0x3e04042c, 0x3e04043e, 0x3e040c04, 0x3e041c14, + 0x3e042c14, 0x3e0c1434, 0x3e0c2404, 0x3e140c14, 0x3e14242c, 0x3e142c14, 0x3e1c0404, 0x3e1c0c2c, + 0x3e1c1c1c, 0x3e1c3404, 0x3e24140c, 0x3e24240c, 0x3e2c0404, 0x3e2c0414, 0x3e2c1424, 0x3e341c04, +GGML_TABLE_END() + +GGML_TABLE_BEGIN(uint32_t, iq3s_grid, 512) + 0x01010101, 0x01010103, 0x01010105, 0x0101010b, 0x0101010f, 0x01010301, 0x01010303, 0x01010305, + 0x01010309, 0x0101030d, 0x01010501, 0x01010503, 0x0101050b, 0x01010707, 0x01010901, 0x01010905, + 0x0101090b, 0x0101090f, 0x01010b03, 0x01010b07, 0x01010d01, 0x01010d05, 0x01010f03, 0x01010f09, + 0x01010f0f, 0x01030101, 0x01030103, 0x01030105, 0x01030109, 0x01030301, 0x01030303, 0x0103030b, + 0x01030501, 0x01030507, 0x0103050f, 0x01030703, 0x0103070b, 0x01030909, 0x01030d03, 0x01030d0b, + 0x01030f05, 0x01050101, 0x01050103, 0x0105010b, 0x0105010f, 0x01050301, 0x01050307, 0x0105030d, + 0x01050503, 0x0105050b, 0x01050701, 0x01050709, 0x01050905, 0x0105090b, 0x0105090f, 0x01050b03, + 0x01050b07, 0x01050f01, 0x01050f07, 0x01070107, 0x01070303, 0x0107030b, 0x01070501, 0x01070505, + 0x01070703, 0x01070707, 0x0107070d, 0x01070909, 0x01070b01, 0x01070b05, 0x01070d0f, 0x01070f03, + 0x01070f0b, 0x01090101, 0x01090307, 0x0109030f, 0x01090503, 0x01090509, 0x01090705, 0x01090901, + 0x01090907, 0x01090b03, 0x01090f01, 0x010b0105, 0x010b0109, 0x010b0501, 0x010b0505, 0x010b050d, + 0x010b0707, 0x010b0903, 0x010b090b, 0x010b090f, 0x010b0d0d, 0x010b0f07, 0x010d010d, 0x010d0303, + 0x010d0307, 0x010d0703, 0x010d0b05, 0x010d0f03, 0x010f0101, 0x010f0105, 0x010f0109, 0x010f0501, + 0x010f0505, 0x010f050d, 0x010f0707, 0x010f0b01, 0x010f0b09, 0x03010101, 0x03010103, 0x03010105, + 0x03010109, 0x03010301, 0x03010303, 0x03010307, 0x0301030b, 0x0301030f, 0x03010501, 0x03010505, + 0x03010703, 0x03010709, 0x0301070d, 0x03010b09, 0x03010b0d, 0x03010d03, 0x03010f05, 0x03030101, + 0x03030103, 0x03030107, 0x0303010d, 0x03030301, 0x03030309, 0x03030503, 0x03030701, 0x03030707, + 0x03030903, 0x03030b01, 0x03030b05, 0x03030f01, 0x03030f0d, 0x03050101, 0x03050305, 0x0305030b, + 0x0305030f, 0x03050501, 0x03050509, 0x03050705, 0x03050901, 0x03050907, 0x03050b0b, 0x03050d01, + 0x03050f05, 0x03070103, 0x03070109, 0x0307010f, 0x03070301, 0x03070307, 0x03070503, 0x0307050f, + 0x03070701, 0x03070709, 0x03070903, 0x03070d05, 0x03070f01, 0x03090107, 0x0309010b, 0x03090305, + 0x03090309, 0x03090703, 0x03090707, 0x03090905, 0x0309090d, 0x03090b01, 0x03090b09, 0x030b0103, + 0x030b0301, 0x030b0307, 0x030b0503, 0x030b0701, 0x030b0705, 0x030b0b03, 0x030d0501, 0x030d0509, + 0x030d050f, 0x030d0909, 0x030d090d, 0x030f0103, 0x030f0107, 0x030f0301, 0x030f0305, 0x030f0503, + 0x030f070b, 0x030f0903, 0x030f0d05, 0x030f0f01, 0x05010101, 0x05010103, 0x05010107, 0x0501010b, + 0x0501010f, 0x05010301, 0x05010305, 0x05010309, 0x0501030d, 0x05010503, 0x05010507, 0x0501050f, + 0x05010701, 0x05010705, 0x05010903, 0x05010907, 0x0501090b, 0x05010b01, 0x05010b05, 0x05010d0f, + 0x05010f01, 0x05010f07, 0x05010f0b, 0x05030101, 0x05030105, 0x05030301, 0x05030307, 0x0503030f, + 0x05030505, 0x0503050b, 0x05030703, 0x05030709, 0x05030905, 0x05030b03, 0x05050103, 0x05050109, + 0x0505010f, 0x05050503, 0x05050507, 0x05050701, 0x0505070f, 0x05050903, 0x05050b07, 0x05050b0f, + 0x05050f03, 0x05050f09, 0x05070101, 0x05070105, 0x0507010b, 0x05070303, 0x05070505, 0x05070509, + 0x05070703, 0x05070707, 0x05070905, 0x05070b01, 0x05070d0d, 0x05090103, 0x0509010f, 0x05090501, + 0x05090507, 0x05090705, 0x0509070b, 0x05090903, 0x05090f05, 0x05090f0b, 0x050b0109, 0x050b0303, + 0x050b0505, 0x050b070f, 0x050b0901, 0x050b0b07, 0x050b0f01, 0x050d0101, 0x050d0105, 0x050d010f, + 0x050d0503, 0x050d0b0b, 0x050d0d03, 0x050f010b, 0x050f0303, 0x050f050d, 0x050f0701, 0x050f0907, + 0x050f0b01, 0x07010105, 0x07010303, 0x07010307, 0x0701030b, 0x0701030f, 0x07010505, 0x07010703, + 0x07010707, 0x0701070b, 0x07010905, 0x07010909, 0x0701090f, 0x07010b03, 0x07010d07, 0x07010f03, + 0x07030103, 0x07030107, 0x0703010b, 0x07030309, 0x07030503, 0x07030507, 0x07030901, 0x07030d01, + 0x07030f05, 0x07030f0d, 0x07050101, 0x07050305, 0x07050501, 0x07050705, 0x07050709, 0x07050b01, + 0x07070103, 0x07070301, 0x07070309, 0x07070503, 0x07070507, 0x0707050f, 0x07070701, 0x07070903, + 0x07070907, 0x0707090f, 0x07070b0b, 0x07070f07, 0x07090107, 0x07090303, 0x0709030d, 0x07090505, + 0x07090703, 0x07090b05, 0x07090d01, 0x07090d09, 0x070b0103, 0x070b0301, 0x070b0305, 0x070b050b, + 0x070b0705, 0x070b0909, 0x070b0b0d, 0x070b0f07, 0x070d030d, 0x070d0903, 0x070f0103, 0x070f0107, + 0x070f0501, 0x070f0505, 0x070f070b, 0x09010101, 0x09010109, 0x09010305, 0x09010501, 0x09010509, + 0x0901050f, 0x09010705, 0x09010903, 0x09010b01, 0x09010f01, 0x09030105, 0x0903010f, 0x09030303, + 0x09030307, 0x09030505, 0x09030701, 0x0903070b, 0x09030907, 0x09030b03, 0x09030b0b, 0x09050103, + 0x09050107, 0x09050301, 0x0905030b, 0x09050503, 0x09050707, 0x09050901, 0x09050b0f, 0x09050d05, + 0x09050f01, 0x09070109, 0x09070303, 0x09070307, 0x09070501, 0x09070505, 0x09070703, 0x0907070b, + 0x09090101, 0x09090105, 0x09090509, 0x0909070f, 0x09090901, 0x09090f03, 0x090b010b, 0x090b010f, + 0x090b0503, 0x090b0d05, 0x090d0307, 0x090d0709, 0x090d0d01, 0x090f0301, 0x090f030b, 0x090f0701, + 0x090f0907, 0x090f0b03, 0x0b010105, 0x0b010301, 0x0b010309, 0x0b010505, 0x0b010901, 0x0b010909, + 0x0b01090f, 0x0b010b05, 0x0b010d0d, 0x0b010f09, 0x0b030103, 0x0b030107, 0x0b03010b, 0x0b030305, + 0x0b030503, 0x0b030705, 0x0b030f05, 0x0b050101, 0x0b050303, 0x0b050507, 0x0b050701, 0x0b05070d, + 0x0b050b07, 0x0b070105, 0x0b07010f, 0x0b070301, 0x0b07050f, 0x0b070909, 0x0b070b03, 0x0b070d0b, + 0x0b070f07, 0x0b090103, 0x0b090109, 0x0b090501, 0x0b090705, 0x0b09090d, 0x0b0b0305, 0x0b0b050d, + 0x0b0b0b03, 0x0b0b0b07, 0x0b0d0905, 0x0b0f0105, 0x0b0f0109, 0x0b0f0505, 0x0d010303, 0x0d010307, + 0x0d01030b, 0x0d010703, 0x0d010707, 0x0d010d01, 0x0d030101, 0x0d030501, 0x0d03050f, 0x0d030d09, + 0x0d050305, 0x0d050709, 0x0d050905, 0x0d050b0b, 0x0d050d05, 0x0d050f01, 0x0d070101, 0x0d070309, + 0x0d070503, 0x0d070901, 0x0d09050b, 0x0d090907, 0x0d090d05, 0x0d0b0101, 0x0d0b0107, 0x0d0b0709, + 0x0d0b0d01, 0x0d0d010b, 0x0d0d0901, 0x0d0f0303, 0x0d0f0307, 0x0f010101, 0x0f010109, 0x0f01010f, + 0x0f010501, 0x0f010505, 0x0f01070d, 0x0f010901, 0x0f010b09, 0x0f010d05, 0x0f030105, 0x0f030303, + 0x0f030509, 0x0f030907, 0x0f03090b, 0x0f050103, 0x0f050109, 0x0f050301, 0x0f05030d, 0x0f050503, + 0x0f050701, 0x0f050b03, 0x0f070105, 0x0f070705, 0x0f07070b, 0x0f070b07, 0x0f090103, 0x0f09010b, + 0x0f090307, 0x0f090501, 0x0f090b01, 0x0f0b0505, 0x0f0b0905, 0x0f0d0105, 0x0f0d0703, 0x0f0f0101, +GGML_TABLE_END() + +#define NGRID_IQ1S 2048 +#define IQ1S_DELTA 0.125f +#define IQ1M_DELTA 0.125f +#if defined(GGML_COMMON_IMPL_C) +GGML_TABLE_BEGIN(uint64_t, iq1s_grid, NGRID_IQ1S) + 0xffffffffffffffff, 0xffffffffffffff01, 0xffffffffffff0000, 0xffffffffffff01ff, + 0xffffffffffff0101, 0xffffffffff00ff00, 0xffffffffff000000, 0xffffffffff01ffff, + 0xffffffffff01ff01, 0xffffffffff0101ff, 0xffffffffff010101, 0xffffffff00ff0000, + 0xffffffff0000ff00, 0xffffffff000000ff, 0xffffffff00000001, 0xffffffff00010000, + 0xffffffff01ffffff, 0xffffffff01ffff01, 0xffffffff01ff01ff, 0xffffffff01ff0101, + 0xffffffff01000000, 0xffffffff0101ffff, 0xffffffff0101ff01, 0xffffffff010101ff, + 0xffffffff01010101, 0xffffff00ffff00ff, 0xffffff00ffff0000, 0xffffff00ff00ff00, + 0xffffff00ff0000ff, 0xffffff00ff000001, 0xffffff00ff000100, 0xffffff00ff000101, + 0xffffff00ff010000, 0xffffff0000ffff00, 0xffffff0000ff0001, 0xffffff0000ff0100, + 0xffffff000000ff01, 0xffffff0000000000, 0xffffff0000000101, 0xffffff000001ff00, + 0xffffff00000100ff, 0xffffff0000010001, 0xffffff00000101ff, 0xffffff0001ff0000, + 0xffffff000100ff00, 0xffffff00010000ff, 0xffffff0001000001, 0xffffff0001010000, + 0xffffff01ffffffff, 0xffffff01ffffff01, 0xffffff01ffff01ff, 0xffffff01ffff0101, + 0xffffff01ff000000, 0xffffff01ff01ffff, 0xffffff01ff01ff01, 0xffffff01ff0101ff, + 0xffffff01ff010101, 0xffffff0100ff0000, 0xffffff010000ff00, 0xffffff0100000100, + 0xffffff01000100ff, 0xffffff0100010100, 0xffffff0101ffffff, 0xffffff0101ffff01, + 0xffffff0101ff01ff, 0xffffff0101ff0101, 0xffffff010100ff00, 0xffffff0101000000, + 0xffffff0101000100, 0xffffff010101ffff, 0xffffff010101ff01, 0xffffff01010101ff, + 0xffffff0101010101, 0xffff00ffff00ff00, 0xffff00ffff0000ff, 0xffff00ffff000001, + 0xffff00ffff010000, 0xffff00ff00ffff00, 0xffff00ff00ff0100, 0xffff00ff00000000, + 0xffff00ff00000101, 0xffff00ff000100ff, 0xffff00ff00010000, 0xffff00ff0100ff00, + 0xffff00ff01000100, 0xffff00ff01010000, 0xffff0000ffffff00, 0xffff0000ffff00ff, + 0xffff0000ffff0000, 0xffff0000ffff0001, 0xffff0000ff000000, 0xffff0000ff0001ff, + 0xffff0000ff000101, 0xffff0000ff010100, 0xffff000000ffffff, 0xffff000000ff0000, + 0xffff000000ff0101, 0xffff00000000ffff, 0xffff00000000ff00, 0xffff0000000000ff, + 0xffff000000000000, 0xffff000000000001, 0xffff000000000100, 0xffff00000001ffff, + 0xffff00000001ff01, 0xffff000000010000, 0xffff0000000101ff, 0xffff000000010101, + 0xffff000001ffff00, 0xffff00000100ff00, 0xffff000001000000, 0xffff0000010001ff, + 0xffff000001000101, 0xffff00000101ff00, 0xffff0000010100ff, 0xffff000001010000, + 0xffff000001010001, 0xffff000001010100, 0xffff0001ff0000ff, 0xffff0001ff000100, + 0xffff000100ffff00, 0xffff000100ff00ff, 0xffff00010000ffff, 0xffff00010000ff01, + 0xffff000100000000, 0xffff0001000001ff, 0xffff00010001ffff, 0xffff00010001ff00, + 0xffff000100010001, 0xffff000100010100, 0xffff000101ff0000, 0xffff00010100ff00, + 0xffff0001010000ff, 0xffff000101000100, 0xffff01ffffffffff, 0xffff01ffffffff01, + 0xffff01ffffff01ff, 0xffff01ffffff0101, 0xffff01ffff000000, 0xffff01ffff01ffff, + 0xffff01ffff01ff01, 0xffff01ffff0101ff, 0xffff01ffff010101, 0xffff01ff00ff0000, + 0xffff01ff0000ff00, 0xffff01ff00000001, 0xffff01ff00010000, 0xffff01ff01ffffff, + 0xffff01ff01ffff01, 0xffff01ff01ff01ff, 0xffff01ff01ff0101, 0xffff01ff01000000, + 0xffff01ff0101ffff, 0xffff01ff0101ff01, 0xffff01ff010101ff, 0xffff01ff01010101, + 0xffff0100ffff0000, 0xffff0100ff00ff00, 0xffff0100ff0000ff, 0xffff0100ff000100, + 0xffff0100ff0100ff, 0xffff0100ff010000, 0xffff010000ffff00, 0xffff01000000ffff, + 0xffff01000000ff00, 0xffff010000000000, 0xffff01000001ff00, 0xffff0100000100ff, + 0xffff010000010100, 0xffff01000100ff00, 0xffff0100010000ff, 0xffff010001000001, + 0xffff010001000100, 0xffff010001010000, 0xffff0101ffffffff, 0xffff0101ffffff01, + 0xffff0101ffff01ff, 0xffff0101ffff0101, 0xffff0101ff000000, 0xffff0101ff01ffff, + 0xffff0101ff01ff01, 0xffff0101ff0101ff, 0xffff0101ff010101, 0xffff010100ff0000, + 0xffff01010000ff00, 0xffff010100000100, 0xffff01010001ff00, 0xffff010100010000, + 0xffff010101ffffff, 0xffff010101ffff01, 0xffff010101ff0000, 0xffff010101ff01ff, + 0xffff010101ff0101, 0xffff010101000000, 0xffff01010101ffff, 0xffff01010101ff01, + 0xffff0101010101ff, 0xffff010101010101, 0xff00ffffff00ffff, 0xff00ffffff00ff00, + 0xff00ffffff0000ff, 0xff00ffffff000100, 0xff00ffffff0100ff, 0xff00ffffff010000, + 0xff00ffff00ffff00, 0xff00ffff00ff00ff, 0xff00ffff0000ffff, 0xff00ffff00000000, + 0xff00ffff000001ff, 0xff00ffff0001ff00, 0xff00ffff000100ff, 0xff00ffff00010000, + 0xff00ffff00010100, 0xff00ffff0100ff00, 0xff00ffff010000ff, 0xff00ffff01000001, + 0xff00ffff0101ff00, 0xff00ffff01010000, 0xff00ff00ffffff00, 0xff00ff00ffff00ff, + 0xff00ff00ffff0001, 0xff00ff00ffff0100, 0xff00ff00ff00ffff, 0xff00ff00ff00ff01, + 0xff00ff00ff000000, 0xff00ff00ff0001ff, 0xff00ff00ff01ff00, 0xff00ff00ff0100ff, + 0xff00ff00ff010100, 0xff00ff0000ff0000, 0xff00ff0000ff0101, 0xff00ff000000ffff, + 0xff00ff000000ff00, 0xff00ff000000ff01, 0xff00ff00000000ff, 0xff00ff0000000000, + 0xff00ff0000000001, 0xff00ff0000000100, 0xff00ff000001ffff, 0xff00ff0000010000, + 0xff00ff0001ff00ff, 0xff00ff000100ff01, 0xff00ff0001000000, 0xff00ff000101ff00, + 0xff00ff00010100ff, 0xff00ff01ff00ff00, 0xff00ff01ff0000ff, 0xff00ff01ff000001, + 0xff00ff01ff010000, 0xff00ff0100ffffff, 0xff00ff0100ff0001, 0xff00ff0100ff0100, + 0xff00ff010000ff01, 0xff00ff0100000000, 0xff00ff01000001ff, 0xff00ff0100000101, + 0xff00ff01000100ff, 0xff00ff0100010001, 0xff00ff0101ff0000, 0xff00ff010100ff00, + 0xff00ff01010000ff, 0xff00ff0101000001, 0xff00ff0101010000, 0xff0000ffffffff00, + 0xff0000ffffff0001, 0xff0000ffffff0100, 0xff0000ffff0000ff, 0xff0000ffff000000, + 0xff0000ffff0001ff, 0xff0000ffff000100, 0xff0000ffff01ff00, 0xff0000ffff010001, + 0xff0000ff00ffff00, 0xff0000ff00ff0000, 0xff0000ff00ff0001, 0xff0000ff00ff01ff, + 0xff0000ff00ff0101, 0xff0000ff0000ff00, 0xff0000ff000000ff, 0xff0000ff00000000, + 0xff0000ff00000001, 0xff0000ff00000100, 0xff0000ff0001ff01, 0xff0000ff00010000, + 0xff0000ff000101ff, 0xff0000ff01ff00ff, 0xff0000ff01ff0100, 0xff0000ff0100ffff, + 0xff0000ff010000ff, 0xff0000ff01000000, 0xff0000ff010001ff, 0xff0000ff01000100, + 0xff0000ff01000101, 0xff0000ff0101ff00, 0xff0000ff010100ff, 0xff0000ff01010000, + 0xff0000ff01010100, 0xff000000ffffff01, 0xff000000ffff0000, 0xff000000ffff0101, + 0xff000000ff00ff00, 0xff000000ff0000ff, 0xff000000ff000000, 0xff000000ff000001, + 0xff000000ff000100, 0xff000000ff01ffff, 0xff000000ff01ff01, 0xff000000ff010000, + 0xff000000ff0101ff, 0xff000000ff010101, 0xff00000000ffff00, 0xff00000000ff00ff, + 0xff00000000ff0000, 0xff00000000ff0001, 0xff0000000000ff00, 0xff0000000000ff01, + 0xff000000000000ff, 0xff00000000000000, 0xff00000000000001, 0xff00000000000100, + 0xff00000000000101, 0xff0000000001ff00, 0xff000000000100ff, 0xff00000000010000, + 0xff00000000010001, 0xff00000000010100, 0xff00000001ffffff, 0xff00000001ffff01, + 0xff00000001ff00ff, 0xff00000001ff0000, 0xff00000001ff01ff, 0xff00000001ff0101, + 0xff0000000100ffff, 0xff0000000100ff00, 0xff000000010000ff, 0xff00000001000000, + 0xff00000001000001, 0xff00000001000100, 0xff00000001000101, 0xff0000000101ffff, + 0xff0000000101ff01, 0xff00000001010000, 0xff000001ffffff00, 0xff000001ffff00ff, + 0xff000001ffff0000, 0xff000001ffff0001, 0xff000001ff000000, 0xff000001ff000001, + 0xff000001ff0001ff, 0xff000001ff000101, 0xff000001ff01ff00, 0xff000001ff010001, + 0xff00000100ffffff, 0xff00000100ffff01, 0xff00000100ff00ff, 0xff00000100ff0000, + 0xff00000100ff01ff, 0xff00000100ff0101, 0xff0000010000ff00, 0xff00000100000000, + 0xff00000100000001, 0xff000001000001ff, 0xff00000100000100, 0xff0000010001ff00, + 0xff000001000100ff, 0xff00000100010000, 0xff000001000101ff, 0xff00000100010100, + 0xff00000100010101, 0xff00000101ff0001, 0xff00000101ff0101, 0xff0000010100ff01, + 0xff00000101000000, 0xff000001010100ff, 0xff00000101010100, 0xff0001ffff00ff00, + 0xff0001ffff000001, 0xff0001ffff010000, 0xff0001ff00ffff00, 0xff0001ff00ff00ff, + 0xff0001ff00ff0001, 0xff0001ff00ff0100, 0xff0001ff0000ffff, 0xff0001ff00000000, + 0xff0001ff000001ff, 0xff0001ff00000101, 0xff0001ff0001ffff, 0xff0001ff0001ff00, + 0xff0001ff000100ff, 0xff0001ff00010001, 0xff0001ff00010100, 0xff0001ff01ff0000, + 0xff0001ff0100ff00, 0xff0001ff010000ff, 0xff0001ff01010000, 0xff000100ff00ffff, + 0xff000100ff00ff01, 0xff000100ff000000, 0xff000100ff000101, 0xff000100ff01ff00, + 0xff000100ff010000, 0xff00010000ffff01, 0xff00010000ff00ff, 0xff00010000ff0000, + 0xff00010000ff01ff, 0xff0001000000ff00, 0xff000100000000ff, 0xff00010000000000, + 0xff00010000000001, 0xff00010000000100, 0xff00010000000101, 0xff0001000001ffff, + 0xff00010000010000, 0xff00010000010101, 0xff00010001ff0100, 0xff0001000100ff00, + 0xff0001000100ff01, 0xff00010001000000, 0xff000100010001ff, 0xff0001000101ff00, + 0xff00010001010001, 0xff00010001010100, 0xff000101ffff0100, 0xff000101ff000001, + 0xff000101ff0100ff, 0xff000101ff010001, 0xff00010100ff00ff, 0xff00010100ff0001, + 0xff00010100ff0100, 0xff0001010000ffff, 0xff0001010000ff01, 0xff00010100000000, + 0xff000101000001ff, 0xff0001010001ff00, 0xff00010100010001, 0xff00010100010100, + 0xff00010101ff0000, 0xff0001010100ff00, 0xff00010101000001, 0xff00010101000101, + 0xff01ffffffffffff, 0xff01ffffffffff01, 0xff01ffffffff01ff, 0xff01ffffffff0101, + 0xff01ffffff000000, 0xff01ffffff01ffff, 0xff01ffffff01ff01, 0xff01ffffff010000, + 0xff01ffffff0101ff, 0xff01ffffff010101, 0xff01ffff00ff0000, 0xff01ffff0000ff00, + 0xff01ffff00000100, 0xff01ffff0001ff00, 0xff01ffff00010000, 0xff01ffff01ffffff, + 0xff01ffff01ffff01, 0xff01ffff01ff01ff, 0xff01ffff01ff0101, 0xff01ffff01000000, + 0xff01ffff0101ffff, 0xff01ffff0101ff01, 0xff01ffff01010000, 0xff01ffff010101ff, + 0xff01ffff01010101, 0xff01ff00ffff0000, 0xff01ff00ff00ff00, 0xff01ff00ff0000ff, + 0xff01ff00ff000100, 0xff01ff00ff010000, 0xff01ff0000ffff01, 0xff01ff0000ff00ff, + 0xff01ff0000ff0100, 0xff01ff0000000000, 0xff01ff00000001ff, 0xff01ff0000000101, + 0xff01ff000001ff00, 0xff01ff00000100ff, 0xff01ff0000010000, 0xff01ff0000010001, + 0xff01ff0001ff0000, 0xff01ff000100ffff, 0xff01ff0001000001, 0xff01ff0001000100, + 0xff01ff0001010000, 0xff01ff01ffffff00, 0xff01ff01ffff01ff, 0xff01ff01ffff0101, + 0xff01ff01ff00ff00, 0xff01ff01ff000000, 0xff01ff01ff01ffff, 0xff01ff01ff01ff01, + 0xff01ff01ff0101ff, 0xff01ff01ff010101, 0xff01ff0100ff0000, 0xff01ff010000ff00, + 0xff01ff0100000001, 0xff01ff0100000100, 0xff01ff0100010000, 0xff01ff0101ffff00, + 0xff01ff0101ff01ff, 0xff01ff0101ff0101, 0xff01ff010100ff00, 0xff01ff0101000000, + 0xff01ff010101ffff, 0xff01ff010101ff01, 0xff01ff01010101ff, 0xff01ff0101010101, + 0xff0100ffffff0000, 0xff0100ffff0000ff, 0xff0100ffff000001, 0xff0100ffff000100, + 0xff0100ffff010000, 0xff0100ff00ff00ff, 0xff0100ff00ff0000, 0xff0100ff00ff0001, + 0xff0100ff00ff0100, 0xff0100ff0000ff01, 0xff0100ff00000000, 0xff0100ff000001ff, + 0xff0100ff00000101, 0xff0100ff00010001, 0xff0100ff01ff0000, 0xff0100ff0100ff00, + 0xff0100ff010000ff, 0xff0100ff01000100, 0xff0100ff0101ff00, 0xff0100ff01010000, + 0xff010000ffff0100, 0xff010000ff000000, 0xff010000ff01ff00, 0xff010000ff010100, + 0xff01000000ffffff, 0xff01000000ff0000, 0xff01000000ff01ff, 0xff0100000000ff00, + 0xff010000000000ff, 0xff01000000000000, 0xff01000000000100, 0xff0100000001ff01, + 0xff01000000010000, 0xff010000000101ff, 0xff01000001ff0100, 0xff0100000100ffff, + 0xff010000010000ff, 0xff01000001000000, 0xff010000010001ff, 0xff01000001000101, + 0xff0100000101ff00, 0xff010000010100ff, 0xff01000001010001, 0xff01000001010100, + 0xff010001ffff0000, 0xff010001ff00ffff, 0xff010001ff00ff01, 0xff010001ff000100, + 0xff010001ff010000, 0xff01000100ffff00, 0xff01000100ff0100, 0xff01000100000000, + 0xff0100010001ffff, 0xff0100010001ff00, 0xff01000100010100, 0xff01000101ff00ff, + 0xff01000101ff0001, 0xff0100010100ffff, 0xff01000101000101, 0xff0101ffffffffff, + 0xff0101ffffffff01, 0xff0101ffffff01ff, 0xff0101ffffff0101, 0xff0101ffff000000, + 0xff0101ffff01ffff, 0xff0101ffff01ff01, 0xff0101ffff0101ff, 0xff0101ffff010101, + 0xff0101ff00ff0000, 0xff0101ff0000ff00, 0xff0101ff000000ff, 0xff0101ff00010000, + 0xff0101ff01ffffff, 0xff0101ff01ffff01, 0xff0101ff01ff01ff, 0xff0101ff01ff0101, + 0xff0101ff0101ffff, 0xff0101ff0101ff01, 0xff0101ff010101ff, 0xff0101ff01010101, + 0xff010100ffff0100, 0xff010100ff00ff00, 0xff010100ff0000ff, 0xff010100ff000100, + 0xff010100ff010000, 0xff01010000ff0001, 0xff01010000ff0100, 0xff0101000000ff01, + 0xff01010000000000, 0xff0101000001ff00, 0xff010100000100ff, 0xff01010000010001, + 0xff01010000010100, 0xff01010001ff0000, 0xff0101000100ffff, 0xff01010001000001, + 0xff01010001000100, 0xff010100010100ff, 0xff01010001010000, 0xff010101ffffffff, + 0xff010101ffffff01, 0xff010101ffff01ff, 0xff010101ffff0101, 0xff010101ff01ffff, + 0xff010101ff01ff01, 0xff010101ff0101ff, 0xff010101ff010101, 0xff01010100ff0000, + 0xff0101010000ff00, 0xff01010100000001, 0xff01010100000100, 0xff01010100010000, + 0xff01010101ffffff, 0xff01010101ffff01, 0xff01010101ff01ff, 0xff01010101ff0101, + 0xff01010101000000, 0xff0101010101ffff, 0xff0101010101ff01, 0xff010101010101ff, + 0xff01010101010101, 0x00ffffffffff0000, 0x00ffffffff00ff00, 0x00ffffffff000001, + 0x00ffffffff010000, 0x00ffffff00ff0100, 0x00ffffff0000ff01, 0x00ffffff00000000, + 0x00ffffff000001ff, 0x00ffffff00000101, 0x00ffffff0001ff00, 0x00ffffff000100ff, + 0x00ffffff00010001, 0x00ffffff010000ff, 0x00ffffff01000100, 0x00ffffff0101ff00, + 0x00ffffff01010001, 0x00ffff00ffffffff, 0x00ffff00ffffff00, 0x00ffff00ffff00ff, + 0x00ffff00ffff0001, 0x00ffff00ffff0100, 0x00ffff00ff00ff01, 0x00ffff00ff000000, + 0x00ffff00ff000001, 0x00ffff00ff0001ff, 0x00ffff00ff000101, 0x00ffff00ff01ff00, + 0x00ffff00ff010001, 0x00ffff00ff010100, 0x00ffff0000ff0000, 0x00ffff0000ff01ff, + 0x00ffff0000ff0101, 0x00ffff000000ff00, 0x00ffff00000000ff, 0x00ffff0000000000, + 0x00ffff0000000001, 0x00ffff0000000100, 0x00ffff0000000101, 0x00ffff0000010000, + 0x00ffff00000101ff, 0x00ffff0000010101, 0x00ffff0001ffff00, 0x00ffff0001ff00ff, + 0x00ffff0001ff0001, 0x00ffff000100ffff, 0x00ffff000100ff01, 0x00ffff0001000000, + 0x00ffff000101ffff, 0x00ffff000101ff00, 0x00ffff000101ff01, 0x00ffff01ffff0000, + 0x00ffff01ff00ff00, 0x00ffff01ff0000ff, 0x00ffff01ff000001, 0x00ffff01ff010000, + 0x00ffff0100ffff00, 0x00ffff010000ff01, 0x00ffff0100000000, 0x00ffff0100000101, + 0x00ffff01000100ff, 0x00ffff0100010100, 0x00ffff0101ff0100, 0x00ffff01010000ff, + 0x00ffff0101010000, 0x00ff00ffffffff00, 0x00ff00ffff000000, 0x00ff00ffff000100, + 0x00ff00ffff010100, 0x00ff00ff00ff0000, 0x00ff00ff00ff01ff, 0x00ff00ff00ff0101, + 0x00ff00ff0000ff00, 0x00ff00ff000000ff, 0x00ff00ff00000000, 0x00ff00ff00000001, + 0x00ff00ff0001ff00, 0x00ff00ff0001ff01, 0x00ff00ff00010000, 0x00ff00ff000101ff, + 0x00ff00ff00010101, 0x00ff00ff01ffff00, 0x00ff00ff01ff0001, 0x00ff00ff01ff0100, + 0x00ff00ff0100ffff, 0x00ff00ff0100ff01, 0x00ff00ff01000000, 0x00ff00ff0101ffff, + 0x00ff00ff0101ff00, 0x00ff00ff01010100, 0x00ff0000ffffff00, 0x00ff0000ffffff01, + 0x00ff0000ffff0000, 0x00ff0000ffff0101, 0x00ff0000ff00ff00, 0x00ff0000ff0000ff, + 0x00ff0000ff000000, 0x00ff0000ff000001, 0x00ff0000ff000100, 0x00ff0000ff01ffff, + 0x00ff0000ff010000, 0x00ff0000ff010101, 0x00ff000000ffff00, 0x00ff000000ff00ff, + 0x00ff000000ff0000, 0x00ff000000ff0001, 0x00ff000000ff0100, 0x00ff00000000ffff, + 0x00ff00000000ff00, 0x00ff0000000000ff, 0x00ff000000000000, 0x00ff000000000001, + 0x00ff0000000001ff, 0x00ff000000000100, 0x00ff00000001ff00, 0x00ff0000000100ff, + 0x00ff000000010000, 0x00ff000000010001, 0x00ff000000010100, 0x00ff000001ffff01, + 0x00ff000001ff00ff, 0x00ff000001ff0000, 0x00ff000001ff01ff, 0x00ff00000100ff00, + 0x00ff0000010000ff, 0x00ff000001000000, 0x00ff000001000001, 0x00ff000001000100, + 0x00ff000001000101, 0x00ff000001010000, 0x00ff0000010101ff, 0x00ff000001010101, + 0x00ff0001ffffff00, 0x00ff0001ffff0000, 0x00ff0001ffff0100, 0x00ff0001ff0000ff, + 0x00ff0001ff000000, 0x00ff0001ff0001ff, 0x00ff0001ff000101, 0x00ff0001ff01ff00, + 0x00ff0001ff0100ff, 0x00ff0001ff010100, 0x00ff000100ffffff, 0x00ff000100ffff01, + 0x00ff000100ff0000, 0x00ff000100ff01ff, 0x00ff00010000ffff, 0x00ff00010000ff00, + 0x00ff00010000ff01, 0x00ff000100000000, 0x00ff000100000001, 0x00ff000100000100, + 0x00ff00010001ff01, 0x00ff000100010000, 0x00ff0001000101ff, 0x00ff000101ffff00, + 0x00ff000101ff0000, 0x00ff000101ff0101, 0x00ff0001010000ff, 0x00ff000101000000, + 0x00ff00010101ff00, 0x00ff0001010100ff, 0x00ff000101010001, 0x00ff01ffffff0000, + 0x00ff01ffff00ff00, 0x00ff01ffff000000, 0x00ff01ffff000101, 0x00ff01ffff010000, + 0x00ff01ff00ffff01, 0x00ff01ff00ff0100, 0x00ff01ff0000ffff, 0x00ff01ff00000000, + 0x00ff01ff000001ff, 0x00ff01ff0001ff00, 0x00ff01ff000100ff, 0x00ff01ff00010001, + 0x00ff01ff00010100, 0x00ff01ff01ff0000, 0x00ff01ff0100ff00, 0x00ff01ff010000ff, + 0x00ff01ff01000001, 0x00ff01ff01000100, 0x00ff01ff01010000, 0x00ff0100ffffff00, + 0x00ff0100ffff0000, 0x00ff0100ffff0001, 0x00ff0100ffff0101, 0x00ff0100ff00ffff, + 0x00ff0100ff0000ff, 0x00ff0100ff000000, 0x00ff0100ff0001ff, 0x00ff0100ff01ff00, + 0x00ff0100ff0100ff, 0x00ff0100ff010001, 0x00ff010000ffffff, 0x00ff010000ff0000, + 0x00ff010000ff0101, 0x00ff01000000ff00, 0x00ff01000000ff01, 0x00ff0100000000ff, + 0x00ff010000000000, 0x00ff010000000001, 0x00ff010000000100, 0x00ff01000001ffff, + 0x00ff01000001ff01, 0x00ff010000010000, 0x00ff010000010001, 0x00ff010000010101, + 0x00ff010001ff0001, 0x00ff010001ff0100, 0x00ff01000100ff01, 0x00ff010001000000, + 0x00ff010001000001, 0x00ff0100010001ff, 0x00ff01000101ff00, 0x00ff0100010100ff, + 0x00ff010001010001, 0x00ff010001010100, 0x00ff0101ff000001, 0x00ff010100ff00ff, + 0x00ff010100ff0001, 0x00ff010100ff0100, 0x00ff010100000000, 0x00ff0101000001ff, + 0x00ff010100000101, 0x00ff0101000100ff, 0x00ff010100010100, 0x00ff0101010000ff, + 0x00ff010101010000, 0x0000ffffffffff00, 0x0000ffffffff00ff, 0x0000ffffffff0000, + 0x0000ffffffff0001, 0x0000ffffffff0100, 0x0000ffffff00ff01, 0x0000ffffff000000, + 0x0000ffffff000101, 0x0000ffffff01ff00, 0x0000ffffff0100ff, 0x0000ffffff010100, + 0x0000ffff00ffffff, 0x0000ffff00ff0000, 0x0000ffff00ff01ff, 0x0000ffff0000ff00, + 0x0000ffff000000ff, 0x0000ffff00000000, 0x0000ffff00000001, 0x0000ffff00000100, + 0x0000ffff00010000, 0x0000ffff000101ff, 0x0000ffff01ff0001, 0x0000ffff01ff0100, + 0x0000ffff01000000, 0x0000ffff010001ff, 0x0000ffff0101ffff, 0x0000ffff0101ff00, + 0x0000ffff01010001, 0x0000ffff01010100, 0x0000ff00ffff0000, 0x0000ff00ffff01ff, + 0x0000ff00ffff0100, 0x0000ff00ffff0101, 0x0000ff00ff00ff00, 0x0000ff00ff0000ff, + 0x0000ff00ff000000, 0x0000ff00ff000001, 0x0000ff00ff0001ff, 0x0000ff00ff000100, + 0x0000ff00ff01ffff, 0x0000ff00ff010000, 0x0000ff00ff010001, 0x0000ff00ff0101ff, + 0x0000ff00ff010101, 0x0000ff0000ffff00, 0x0000ff0000ff00ff, 0x0000ff0000ff0000, + 0x0000ff0000ff0001, 0x0000ff0000ff0100, 0x0000ff000000ffff, 0x0000ff000000ff00, + 0x0000ff000000ff01, 0x0000ff00000000ff, 0x0000ff0000000000, 0x0000ff0000000001, + 0x0000ff00000001ff, 0x0000ff0000000100, 0x0000ff0000000101, 0x0000ff000001ff00, + 0x0000ff00000100ff, 0x0000ff0000010000, 0x0000ff0000010001, 0x0000ff0000010100, + 0x0000ff0001ffff01, 0x0000ff0001ff0000, 0x0000ff000100ff00, 0x0000ff00010000ff, + 0x0000ff0001000000, 0x0000ff0001000001, 0x0000ff0001000100, 0x0000ff000101ffff, + 0x0000ff0001010000, 0x0000ff0001010101, 0x0000ff01ffffff00, 0x0000ff01ffff0001, + 0x0000ff01ff00ff01, 0x0000ff01ff000000, 0x0000ff01ff000101, 0x0000ff01ff01ff00, + 0x0000ff01ff0100ff, 0x0000ff0100ffff01, 0x0000ff0100ff0000, 0x0000ff0100ff0101, + 0x0000ff010000ff00, 0x0000ff01000000ff, 0x0000ff0100000000, 0x0000ff0100000001, + 0x0000ff0100000100, 0x0000ff010001ff01, 0x0000ff0100010000, 0x0000ff0101ff0000, + 0x0000ff010100ffff, 0x0000ff010100ff01, 0x0000ff0101000000, 0x0000ff0101000100, + 0x0000ff0101000101, 0x0000ff01010100ff, 0x000000ffffff00ff, 0x000000ffffff0000, + 0x000000ffff00ff00, 0x000000ffff0000ff, 0x000000ffff000000, 0x000000ffff000001, + 0x000000ffff0001ff, 0x000000ffff000100, 0x000000ffff01ff00, 0x000000ffff010000, + 0x000000ffff0101ff, 0x000000ffff010101, 0x000000ff00ffff00, 0x000000ff00ff00ff, + 0x000000ff00ff0000, 0x000000ff00ff0001, 0x000000ff00ff0100, 0x000000ff00ff0101, + 0x000000ff0000ffff, 0x000000ff0000ff00, 0x000000ff000000ff, 0x000000ff00000000, + 0x000000ff00000001, 0x000000ff000001ff, 0x000000ff00000100, 0x000000ff00000101, + 0x000000ff0001ff00, 0x000000ff0001ff01, 0x000000ff000100ff, 0x000000ff00010000, + 0x000000ff00010001, 0x000000ff00010100, 0x000000ff01ffffff, 0x000000ff01ff01ff, + 0x000000ff01ff0101, 0x000000ff0100ff00, 0x000000ff010000ff, 0x000000ff01000000, + 0x000000ff01000001, 0x000000ff01000100, 0x000000ff0101ff00, 0x000000ff010100ff, + 0x000000ff01010000, 0x000000ff01010101, 0x00000000ffffff00, 0x00000000ffffff01, + 0x00000000ffff00ff, 0x00000000ffff0000, 0x00000000ffff0001, 0x00000000ffff0100, + 0x00000000ff00ffff, 0x00000000ff00ff00, 0x00000000ff00ff01, 0x00000000ff0000ff, + 0x00000000ff000000, 0x00000000ff000001, 0x00000000ff000100, 0x00000000ff000101, + 0x00000000ff01ff00, 0x00000000ff0100ff, 0x00000000ff010000, 0x00000000ff010001, + 0x00000000ff010100, 0x0000000000ffffff, 0x0000000000ffff00, 0x0000000000ffff01, + 0x0000000000ff00ff, 0x0000000000ff0000, 0x0000000000ff0001, 0x0000000000ff01ff, + 0x0000000000ff0100, 0x000000000000ffff, 0x000000000000ff00, 0x000000000000ff01, + 0x00000000000000ff, 0x0000000000000000, 0x0000000000000001, 0x00000000000001ff, + 0x0000000000000100, 0x0000000000000101, 0x000000000001ffff, 0x000000000001ff00, + 0x00000000000100ff, 0x0000000000010000, 0x0000000000010001, 0x00000000000101ff, + 0x0000000000010100, 0x0000000000010101, 0x0000000001ffff00, 0x0000000001ff00ff, + 0x0000000001ff0000, 0x0000000001ff0100, 0x0000000001ff0101, 0x000000000100ffff, + 0x000000000100ff00, 0x00000000010000ff, 0x0000000001000000, 0x0000000001000001, + 0x00000000010001ff, 0x0000000001000100, 0x000000000101ff00, 0x00000000010100ff, + 0x0000000001010000, 0x0000000001010001, 0x0000000001010100, 0x00000001ffffffff, + 0x00000001ffffff00, 0x00000001ffffff01, 0x00000001ffff00ff, 0x00000001ffff0001, + 0x00000001ffff01ff, 0x00000001ffff0100, 0x00000001ff00ff00, 0x00000001ff0000ff, + 0x00000001ff000000, 0x00000001ff0001ff, 0x00000001ff000100, 0x00000001ff01ffff, + 0x00000001ff01ff00, 0x00000001ff01ff01, 0x00000001ff0100ff, 0x00000001ff010000, + 0x00000001ff010001, 0x00000001ff0101ff, 0x00000001ff010100, 0x0000000100ffff00, + 0x0000000100ff0000, 0x0000000100ff0001, 0x0000000100ff01ff, 0x0000000100ff0100, + 0x0000000100ff0101, 0x000000010000ffff, 0x000000010000ff00, 0x000000010000ff01, + 0x00000001000000ff, 0x0000000100000000, 0x0000000100000001, 0x00000001000001ff, + 0x0000000100000100, 0x0000000100000101, 0x000000010001ff00, 0x00000001000100ff, + 0x0000000100010000, 0x0000000100010100, 0x0000000101ffff01, 0x0000000101ff0000, + 0x0000000101ff0001, 0x0000000101ff01ff, 0x0000000101ff0100, 0x0000000101ff0101, + 0x000000010100ff00, 0x0000000101000000, 0x0000000101000101, 0x000000010101ff01, + 0x0000000101010000, 0x0000000101010001, 0x00000001010101ff, 0x0000000101010100, + 0x000001ffffff00ff, 0x000001ffffff0000, 0x000001ffffff0001, 0x000001ffffff0100, + 0x000001ffff00ffff, 0x000001ffff000000, 0x000001ffff0001ff, 0x000001ffff01ff00, + 0x000001ffff010101, 0x000001ff00ff0000, 0x000001ff00ff01ff, 0x000001ff00ff0101, + 0x000001ff0000ff00, 0x000001ff000000ff, 0x000001ff00000000, 0x000001ff00000001, + 0x000001ff000001ff, 0x000001ff00000100, 0x000001ff0001ffff, 0x000001ff0001ff01, + 0x000001ff000100ff, 0x000001ff00010000, 0x000001ff01ffff01, 0x000001ff01ff0100, + 0x000001ff0100ffff, 0x000001ff0100ff01, 0x000001ff01000000, 0x000001ff010001ff, + 0x000001ff0101ff00, 0x000001ff01010100, 0x00000100ffffff00, 0x00000100ffffff01, + 0x00000100ffff0000, 0x00000100ffff0101, 0x00000100ff00ff00, 0x00000100ff0000ff, + 0x00000100ff000000, 0x00000100ff000001, 0x00000100ff000100, 0x00000100ff010000, + 0x0000010000ffff00, 0x0000010000ff00ff, 0x0000010000ff0000, 0x0000010000ff0001, + 0x0000010000ff0100, 0x000001000000ffff, 0x000001000000ff00, 0x000001000000ff01, + 0x00000100000000ff, 0x0000010000000000, 0x0000010000000001, 0x00000100000001ff, + 0x0000010000000100, 0x0000010000000101, 0x000001000001ff00, 0x00000100000100ff, + 0x0000010000010000, 0x0000010000010001, 0x0000010000010100, 0x0000010001ffff00, + 0x0000010001ff0000, 0x0000010001ff0100, 0x000001000100ff00, 0x00000100010000ff, + 0x0000010001000000, 0x0000010001000001, 0x00000100010001ff, 0x0000010001000100, + 0x0000010001010000, 0x00000101ffff00ff, 0x00000101ffff01ff, 0x00000101ff000000, + 0x00000101ff000101, 0x00000101ff01ffff, 0x00000101ff010000, 0x00000101ff010001, + 0x00000101ff010100, 0x0000010100ff0000, 0x0000010100ff01ff, 0x0000010100ff0100, + 0x000001010000ff00, 0x0000010100000000, 0x0000010100000001, 0x00000101000001ff, + 0x0000010100000100, 0x000001010001ff01, 0x0000010100010000, 0x00000101000101ff, + 0x0000010100010101, 0x0000010101ffff00, 0x0000010101ff0101, 0x000001010100ff01, + 0x0000010101000000, 0x0000010101000001, 0x00000101010001ff, 0x0000010101000101, + 0x000001010101ff00, 0x0001ffffffff0000, 0x0001ffffff0000ff, 0x0001ffffff000001, + 0x0001ffffff000100, 0x0001ffffff010000, 0x0001ffff00ff00ff, 0x0001ffff0000ffff, + 0x0001ffff00000000, 0x0001ffff00000001, 0x0001ffff000001ff, 0x0001ffff00000101, + 0x0001ffff0001ff00, 0x0001ffff000100ff, 0x0001ffff00010001, 0x0001ffff00010100, + 0x0001ffff01ffff00, 0x0001ffff01000001, 0x0001ffff01010000, 0x0001ff00ffffff00, + 0x0001ff00ffff00ff, 0x0001ff00ffff0001, 0x0001ff00ffff0100, 0x0001ff00ff00ff01, + 0x0001ff00ff000000, 0x0001ff00ff01ff00, 0x0001ff00ff01ff01, 0x0001ff00ff010001, + 0x0001ff00ff010100, 0x0001ff0000ff0000, 0x0001ff0000ff0100, 0x0001ff000000ff00, + 0x0001ff0000000000, 0x0001ff0000000001, 0x0001ff0000000100, 0x0001ff0000010000, + 0x0001ff0000010001, 0x0001ff0000010101, 0x0001ff0001ff00ff, 0x0001ff0001ff0101, + 0x0001ff000100ff01, 0x0001ff0001000000, 0x0001ff000101ff00, 0x0001ff0001010001, + 0x0001ff0001010100, 0x0001ff01ff00ff00, 0x0001ff01ff000001, 0x0001ff01ff000100, + 0x0001ff0100ffffff, 0x0001ff0100ffff00, 0x0001ff0100ff0001, 0x0001ff0100000000, + 0x0001ff0100000001, 0x0001ff01000001ff, 0x0001ff010001ffff, 0x0001ff0101ff0000, + 0x0001ff010100ff00, 0x0001ff0101000001, 0x0001ff0101010000, 0x000100ffff00ff00, + 0x000100ffff00ff01, 0x000100ffff000000, 0x000100ffff000001, 0x000100ffff000101, + 0x000100ffff01ff00, 0x000100ffff010001, 0x000100ffff010100, 0x000100ff00ffffff, + 0x000100ff00ffff01, 0x000100ff00ff0000, 0x000100ff00ff01ff, 0x000100ff00ff0101, + 0x000100ff0000ff00, 0x000100ff000000ff, 0x000100ff00000000, 0x000100ff00000001, + 0x000100ff00000100, 0x000100ff00000101, 0x000100ff0001ffff, 0x000100ff0001ff01, + 0x000100ff00010000, 0x000100ff01ff00ff, 0x000100ff01ff0000, 0x000100ff01ff0100, + 0x000100ff0100ffff, 0x000100ff0100ff01, 0x000100ff010000ff, 0x000100ff01000000, + 0x000100ff01000001, 0x000100ff010001ff, 0x000100ff01000101, 0x000100ff0101ff00, + 0x000100ff010100ff, 0x000100ff01010100, 0x00010000ffff0000, 0x00010000ffff01ff, + 0x00010000ffff0101, 0x00010000ff00ff00, 0x00010000ff000000, 0x00010000ff000001, + 0x00010000ff000100, 0x0001000000ff00ff, 0x0001000000ff0000, 0x0001000000ff0001, + 0x0001000000ff0100, 0x000100000000ffff, 0x000100000000ff00, 0x00010000000000ff, + 0x0001000000000000, 0x0001000000000001, 0x0001000000000100, 0x000100000001ff00, + 0x00010000000100ff, 0x0001000000010000, 0x0001000000010001, 0x0001000000010100, + 0x0001000001ff0001, 0x0001000001ff0100, 0x0001000001ff0101, 0x000100000100ff00, + 0x0001000001000000, 0x0001000001000001, 0x0001000001000100, 0x0001000001000101, + 0x000100000101ff01, 0x0001000001010000, 0x0001000001010001, 0x00010000010101ff, + 0x00010001ffffff01, 0x00010001ffff0100, 0x00010001ff000000, 0x00010001ff01ffff, + 0x00010001ff010001, 0x00010001ff0101ff, 0x00010001ff010100, 0x0001000100ffffff, + 0x0001000100ff0000, 0x0001000100ff01ff, 0x0001000100ff0101, 0x000100010000ff00, + 0x00010001000000ff, 0x0001000100000000, 0x0001000100000001, 0x00010001000001ff, + 0x0001000100000101, 0x000100010001ffff, 0x0001000100010000, 0x00010001000101ff, + 0x0001000101ffffff, 0x0001000101ffff01, 0x0001000101ff0000, 0x0001000101ff0101, + 0x00010001010000ff, 0x0001000101000001, 0x00010001010001ff, 0x0001000101000100, + 0x000100010101ffff, 0x00010001010100ff, 0x0001000101010001, 0x0001000101010101, + 0x000101ffff000001, 0x000101ffff000100, 0x000101ffff010000, 0x000101ff00ffff00, + 0x000101ff0000ff01, 0x000101ff00000000, 0x000101ff00000101, 0x000101ff0001ff00, + 0x000101ff00010100, 0x000101ff01ff0000, 0x000101ff0100ff00, 0x000101ff010001ff, + 0x000101ff01010001, 0x00010100ffffff00, 0x00010100ffff00ff, 0x00010100ff00ffff, + 0x00010100ff000000, 0x00010100ff01ff00, 0x00010100ff0100ff, 0x00010100ff010001, + 0x00010100ff010100, 0x0001010000ffffff, 0x0001010000ffff00, 0x0001010000ff0000, + 0x0001010000ff0001, 0x0001010000ff01ff, 0x000101000000ff00, 0x00010100000000ff, + 0x0001010000000000, 0x0001010000000001, 0x0001010000000100, 0x000101000001ffff, + 0x0001010000010000, 0x0001010000010101, 0x0001010001ffff01, 0x0001010001ff00ff, + 0x0001010001ff0101, 0x0001010001000000, 0x000101000101ff00, 0x00010100010100ff, + 0x0001010001010000, 0x0001010001010100, 0x00010101ff00ff00, 0x00010101ff000001, + 0x00010101ff0001ff, 0x0001010100ffff00, 0x0001010100ff00ff, 0x0001010100ff0100, + 0x000101010000ffff, 0x0001010100000000, 0x00010101000001ff, 0x0001010100000101, + 0x00010101000100ff, 0x0001010100010000, 0x0001010100010100, 0x0001010101ff0001, + 0x00010101010000ff, 0x00010101010001ff, 0x0001010101000101, 0x0001010101010001, + 0x01ffffffffffffff, 0x01ffffffffffff01, 0x01ffffffffff01ff, 0x01ffffffffff0101, + 0x01ffffffff01ffff, 0x01ffffffff01ff01, 0x01ffffffff0101ff, 0x01ffffffff010101, + 0x01ffffff00ff0000, 0x01ffffff0000ffff, 0x01ffffff0000ff00, 0x01ffffff000000ff, + 0x01ffffff00000001, 0x01ffffff00000100, 0x01ffffff00010000, 0x01ffffff01ffffff, + 0x01ffffff01ffff01, 0x01ffffff01ff01ff, 0x01ffffff01ff0101, 0x01ffffff01000000, + 0x01ffffff0101ffff, 0x01ffffff0101ff01, 0x01ffffff010101ff, 0x01ffffff01010101, + 0x01ffff00ffff0000, 0x01ffff00ff00ff00, 0x01ffff00ff0000ff, 0x01ffff00ff000001, + 0x01ffff00ff000100, 0x01ffff00ff010000, 0x01ffff0000ffff00, 0x01ffff0000ff00ff, + 0x01ffff0000ff0100, 0x01ffff000000ffff, 0x01ffff000000ff01, 0x01ffff0000000000, + 0x01ffff0000000001, 0x01ffff00000001ff, 0x01ffff0000000100, 0x01ffff00000100ff, + 0x01ffff0000010001, 0x01ffff0000010100, 0x01ffff0001ff0000, 0x01ffff0001ff0100, + 0x01ffff00010000ff, 0x01ffff0001000001, 0x01ffff0001000100, 0x01ffff0001010000, + 0x01ffff01ffffffff, 0x01ffff01ffffff01, 0x01ffff01ffff01ff, 0x01ffff01ffff0101, + 0x01ffff01ff000000, 0x01ffff01ff01ffff, 0x01ffff01ff01ff01, 0x01ffff01ff0101ff, + 0x01ffff01ff010101, 0x01ffff010000ff00, 0x01ffff01000000ff, 0x01ffff0100000100, + 0x01ffff0100010000, 0x01ffff0101ffffff, 0x01ffff0101ffff01, 0x01ffff0101ff01ff, + 0x01ffff0101ff0101, 0x01ffff0101000000, 0x01ffff010101ffff, 0x01ffff010101ff01, + 0x01ffff01010101ff, 0x01ffff0101010101, 0x01ff00ffff0000ff, 0x01ff00ffff000100, + 0x01ff00ff00ffff00, 0x01ff00ff00ff00ff, 0x01ff00ff0000ff00, 0x01ff00ff00000000, + 0x01ff00ff00000101, 0x01ff00ff0001ff00, 0x01ff00ff000100ff, 0x01ff00ff00010100, + 0x01ff00ff010000ff, 0x01ff00ff01000100, 0x01ff0000ffffff00, 0x01ff0000ffff0100, + 0x01ff0000ff00ff01, 0x01ff0000ff000000, 0x01ff0000ff000101, 0x01ff0000ff010001, + 0x01ff0000ff010100, 0x01ff000000ffffff, 0x01ff000000ffff00, 0x01ff000000ff0000, + 0x01ff000000ff01ff, 0x01ff00000000ff00, 0x01ff0000000000ff, 0x01ff000000000000, + 0x01ff000000000001, 0x01ff000000000100, 0x01ff000000000101, 0x01ff000000010000, + 0x01ff000000010001, 0x01ff0000000101ff, 0x01ff000000010101, 0x01ff000001ffff00, + 0x01ff000001ff00ff, 0x01ff000001ff0001, 0x01ff000001ff0100, 0x01ff00000100ffff, + 0x01ff00000100ff01, 0x01ff000001000000, 0x01ff0000010001ff, 0x01ff000001010001, + 0x01ff0001ff00ff00, 0x01ff0001ff000001, 0x01ff0001ff000100, 0x01ff0001ff010000, + 0x01ff000100ffff00, 0x01ff000100ff00ff, 0x01ff000100ff0100, 0x01ff000100ff0101, + 0x01ff00010000ffff, 0x01ff000100000000, 0x01ff000100000100, 0x01ff000100000101, + 0x01ff00010001ff00, 0x01ff000100010001, 0x01ff000100010101, 0x01ff000101ff0000, + 0x01ff00010100ff00, 0x01ff000101000101, 0x01ff0001010100ff, 0x01ff01ffffffffff, + 0x01ff01ffffffff01, 0x01ff01ffffff01ff, 0x01ff01ffffff0101, 0x01ff01ffff000000, + 0x01ff01ffff01ffff, 0x01ff01ffff01ff01, 0x01ff01ffff0101ff, 0x01ff01ffff010101, + 0x01ff01ff00ffff00, 0x01ff01ff00ff0000, 0x01ff01ff0000ff00, 0x01ff01ff000000ff, + 0x01ff01ff00000100, 0x01ff01ff00010000, 0x01ff01ff00010100, 0x01ff01ff01ffffff, + 0x01ff01ff01ffff01, 0x01ff01ff01ff01ff, 0x01ff01ff01ff0101, 0x01ff01ff01000000, + 0x01ff01ff0101ffff, 0x01ff01ff0101ff01, 0x01ff01ff010101ff, 0x01ff01ff01010101, + 0x01ff0100ffff0000, 0x01ff0100ffff0001, 0x01ff0100ff00ff00, 0x01ff0100ff0000ff, + 0x01ff0100ff000001, 0x01ff0100ff010000, 0x01ff010000ffff00, 0x01ff010000ff00ff, + 0x01ff010000ff0001, 0x01ff010000ff0100, 0x01ff01000000ffff, 0x01ff01000000ff01, + 0x01ff010000000000, 0x01ff010000000101, 0x01ff01000001ff00, 0x01ff0100000100ff, + 0x01ff010001ff0000, 0x01ff010001000001, 0x01ff010001000100, 0x01ff010001010000, + 0x01ff0101ffffffff, 0x01ff0101ffffff01, 0x01ff0101ffff01ff, 0x01ff0101ffff0101, + 0x01ff0101ff000000, 0x01ff0101ff01ffff, 0x01ff0101ff01ff01, 0x01ff0101ff0101ff, + 0x01ff0101ff010101, 0x01ff010100ff0000, 0x01ff01010000ff00, 0x01ff0101000000ff, + 0x01ff010100000001, 0x01ff010101ffffff, 0x01ff010101ffff01, 0x01ff010101ff01ff, + 0x01ff010101ff0101, 0x01ff010101000000, 0x01ff01010101ffff, 0x01ff01010101ff01, + 0x01ff0101010101ff, 0x01ff010101010101, 0x0100ffffffff0000, 0x0100ffffff00ff00, + 0x0100ffffff000001, 0x0100ffffff0001ff, 0x0100ffffff000100, 0x0100ffffff010000, + 0x0100ffff00ffff00, 0x0100ffff00ff0001, 0x0100ffff00ff0100, 0x0100ffff00000000, + 0x0100ffff000001ff, 0x0100ffff00000101, 0x0100ffff00010100, 0x0100ffff00010101, + 0x0100ffff01ff0000, 0x0100ffff0100ff00, 0x0100ffff010000ff, 0x0100ffff01000001, + 0x0100ffff01000100, 0x0100ffff01010000, 0x0100ff00ffffff00, 0x0100ff00ffff00ff, + 0x0100ff00ffff0001, 0x0100ff00ffff0100, 0x0100ff00ff00ffff, 0x0100ff00ff000000, + 0x0100ff00ff0001ff, 0x0100ff00ff000101, 0x0100ff00ff01ff00, 0x0100ff00ff0100ff, + 0x0100ff00ff010001, 0x0100ff00ff010100, 0x0100ff0000ffffff, 0x0100ff0000ff0000, + 0x0100ff000000ffff, 0x0100ff000000ff00, 0x0100ff00000000ff, 0x0100ff0000000000, + 0x0100ff0000000001, 0x0100ff0000000100, 0x0100ff000001ff01, 0x0100ff0000010000, + 0x0100ff0001ff00ff, 0x0100ff0001ff0001, 0x0100ff000100ff01, 0x0100ff0001000000, + 0x0100ff00010001ff, 0x0100ff000101ff00, 0x0100ff00010100ff, 0x0100ff0001010001, + 0x0100ff0001010100, 0x0100ff01ffff0000, 0x0100ff01ff00ff00, 0x0100ff01ff0000ff, + 0x0100ff01ff000100, 0x0100ff01ff010000, 0x0100ff0100ff00ff, 0x0100ff0100ff0001, + 0x0100ff0100ff0100, 0x0100ff010000ffff, 0x0100ff010000ff01, 0x0100ff0100000000, + 0x0100ff01000001ff, 0x0100ff0100010001, 0x0100ff0100010100, 0x0100ff0101ff0000, + 0x0100ff01010000ff, 0x0100ff0101000001, 0x0100ff0101010100, 0x010000ffffffff00, + 0x010000ffffff00ff, 0x010000ffffff0001, 0x010000ffff00ffff, 0x010000ffff000000, + 0x010000ffff0001ff, 0x010000ffff010001, 0x010000ff00ffffff, 0x010000ff00ff0101, + 0x010000ff0000ff00, 0x010000ff000000ff, 0x010000ff00000000, 0x010000ff00000001, + 0x010000ff000001ff, 0x010000ff00000100, 0x010000ff0001ffff, 0x010000ff0001ff00, + 0x010000ff0001ff01, 0x010000ff00010000, 0x010000ff01ff00ff, 0x010000ff01ff0001, + 0x010000ff0100ff01, 0x010000ff010000ff, 0x010000ff01000000, 0x010000ff010001ff, + 0x010000ff0101ff00, 0x010000ff01010100, 0x01000000ffffffff, 0x01000000ffff0000, + 0x01000000ffff01ff, 0x01000000ffff0101, 0x01000000ff00ffff, 0x01000000ff00ff00, + 0x01000000ff0000ff, 0x01000000ff000000, 0x01000000ff000001, 0x01000000ff000100, + 0x01000000ff01ff00, 0x01000000ff010000, 0x01000000ff010100, 0x01000000ff010101, + 0x0100000000ffff00, 0x0100000000ff00ff, 0x0100000000ff0000, 0x0100000000ff0001, + 0x0100000000ff0100, 0x010000000000ffff, 0x010000000000ff00, 0x010000000000ff01, + 0x01000000000000ff, 0x0100000000000000, 0x0100000000000001, 0x01000000000001ff, + 0x0100000000000100, 0x0100000000000101, 0x010000000001ff00, 0x01000000000100ff, + 0x0100000000010000, 0x0100000000010001, 0x0100000000010100, 0x0100000001ffff00, + 0x0100000001ff0000, 0x0100000001ff01ff, 0x010000000100ff00, 0x010000000100ff01, + 0x01000000010000ff, 0x0100000001000000, 0x0100000001000001, 0x0100000001000100, + 0x0100000001000101, 0x010000000101ffff, 0x010000000101ff01, 0x0100000001010000, + 0x01000000010101ff, 0x0100000001010101, 0x01000001ffffff00, 0x01000001ffff00ff, + 0x01000001ff00ffff, 0x01000001ff000000, 0x01000001ff000100, 0x01000001ff01ffff, + 0x01000001ff010001, 0x01000001ff010100, 0x0100000100ff0000, 0x0100000100ff01ff, + 0x0100000100ff0100, 0x010000010000ff00, 0x010000010000ff01, 0x0100000100000000, + 0x0100000100000001, 0x0100000100000100, 0x0100000100010000, 0x01000001000101ff, + 0x0100000101ffff01, 0x0100000101ff00ff, 0x0100000101ff0100, 0x0100000101ff0101, + 0x010000010100ff01, 0x01000001010000ff, 0x0100000101000000, 0x01000001010100ff, + 0x0100000101010001, 0x0100000101010100, 0x010001ffffff0000, 0x010001ffff000001, + 0x010001ffff000100, 0x010001ffff010000, 0x010001ff00ffff00, 0x010001ff00ff0001, + 0x010001ff0000ffff, 0x010001ff0000ff01, 0x010001ff00000000, 0x010001ff00000001, + 0x010001ff00000101, 0x010001ff000100ff, 0x010001ff00010000, 0x010001ff01ff0000, + 0x010001ff0100ff00, 0x010001ff01000001, 0x010001ff01000100, 0x010001ff01010000, + 0x01000100ffff00ff, 0x01000100ffff0001, 0x01000100ffff0100, 0x01000100ff00ffff, + 0x01000100ff00ff01, 0x01000100ff000000, 0x01000100ff0001ff, 0x01000100ff000101, + 0x01000100ff01ffff, 0x01000100ff01ff00, 0x01000100ff0100ff, 0x01000100ff010001, + 0x0100010000ffffff, 0x0100010000ffff01, 0x0100010000ff0000, 0x0100010000ff01ff, + 0x0100010000ff0101, 0x010001000000ff00, 0x01000100000000ff, 0x0100010000000000, + 0x0100010000000001, 0x0100010000000100, 0x010001000001ff01, 0x0100010000010000, + 0x0100010000010001, 0x0100010000010101, 0x0100010001ffff00, 0x0100010001ff00ff, + 0x010001000100ffff, 0x010001000100ff01, 0x0100010001000000, 0x0100010001000101, + 0x010001000101ff00, 0x0100010001010001, 0x01000101ffff0000, 0x01000101ff000000, + 0x01000101ff010000, 0x0100010100ff00ff, 0x0100010100ff0001, 0x0100010100ff0100, + 0x010001010000ffff, 0x0100010100000000, 0x01000101000001ff, 0x010001010001ff00, + 0x0100010101ff0000, 0x010001010100ff00, 0x01000101010000ff, 0x0100010101000000, + 0x0100010101000001, 0x0101ffffffffffff, 0x0101ffffffffff01, 0x0101ffffffff01ff, + 0x0101ffffffff0101, 0x0101ffffff000000, 0x0101ffffff01ffff, 0x0101ffffff01ff01, + 0x0101ffffff0101ff, 0x0101ffffff010101, 0x0101ffff00ff0000, 0x0101ffff0000ff00, + 0x0101ffff000000ff, 0x0101ffff00000001, 0x0101ffff00000100, 0x0101ffff01ffffff, + 0x0101ffff01ffff01, 0x0101ffff01ff01ff, 0x0101ffff01ff0101, 0x0101ffff01000000, + 0x0101ffff0101ffff, 0x0101ffff0101ff01, 0x0101ffff010101ff, 0x0101ffff01010101, + 0x0101ff00ffff0000, 0x0101ff00ffff0100, 0x0101ff00ff00ff00, 0x0101ff00ff0000ff, + 0x0101ff00ff000001, 0x0101ff00ff000100, 0x0101ff00ff000101, 0x0101ff0000ff0001, + 0x0101ff0000ff0100, 0x0101ff000000ff00, 0x0101ff0000000000, 0x0101ff00000001ff, + 0x0101ff0000000101, 0x0101ff000001ff00, 0x0101ff00000100ff, 0x0101ff0001ff0000, + 0x0101ff000100ffff, 0x0101ff000100ff01, 0x0101ff0001000001, 0x0101ff0001000100, + 0x0101ff01ffffff01, 0x0101ff01ffff01ff, 0x0101ff01ffff0101, 0x0101ff01ff00ffff, + 0x0101ff01ff000100, 0x0101ff01ff01ff01, 0x0101ff01ff0101ff, 0x0101ff01ff010101, + 0x0101ff0100ff0000, 0x0101ff010000ff00, 0x0101ff0100000001, 0x0101ff0100000100, + 0x0101ff0100010000, 0x0101ff0101ffffff, 0x0101ff0101ffff01, 0x0101ff0101ff01ff, + 0x0101ff0101ff0101, 0x0101ff0101000000, 0x0101ff010101ffff, 0x0101ff010101ff01, + 0x0101ff01010101ff, 0x0101ff0101010101, 0x010100ffff000100, 0x010100ffff010000, + 0x010100ff00ffff00, 0x010100ff00ff00ff, 0x010100ff0000ffff, 0x010100ff000000ff, + 0x010100ff00000000, 0x010100ff000001ff, 0x010100ff00000101, 0x010100ff0001ff00, + 0x010100ff00010000, 0x010100ff00010001, 0x010100ff000101ff, 0x010100ff00010100, + 0x010100ff01ff0000, 0x01010000ffff0001, 0x01010000ffff0100, 0x01010000ff00ffff, + 0x01010000ff00ff01, 0x01010000ff000000, 0x01010000ff0001ff, 0x01010000ff010001, + 0x01010000ff010100, 0x0101000000ffff01, 0x0101000000ff0000, 0x010100000000ff00, + 0x01010000000000ff, 0x0101000000000000, 0x0101000000000001, 0x0101000000000100, + 0x0101000000010000, 0x0101000000010101, 0x0101000001ffff00, 0x0101000001ff00ff, + 0x0101000001ff0000, 0x0101000001ff0001, 0x0101000001ff0100, 0x010100000100ff01, + 0x0101000001000000, 0x01010000010001ff, 0x01010001ffff0000, 0x01010001ff00ff00, + 0x01010001ff000001, 0x01010001ff000101, 0x01010001ff01ff00, 0x01010001ff010000, + 0x0101000100ff00ff, 0x0101000100ff0001, 0x0101000100ff0101, 0x010100010000ff01, + 0x0101000100000000, 0x0101000100000001, 0x01010001000001ff, 0x010100010001ffff, + 0x010100010001ff01, 0x0101000101ff0001, 0x010100010100ffff, 0x0101000101000000, + 0x0101000101000001, 0x0101000101000100, 0x010100010101ff00, 0x01010001010100ff, + 0x0101000101010001, 0x010101ffffffffff, 0x010101ffffffff01, 0x010101ffffff01ff, + 0x010101ffffff0101, 0x010101ffff01ffff, 0x010101ffff01ff01, 0x010101ffff0101ff, + 0x010101ffff010101, 0x010101ff0000ff00, 0x010101ff000000ff, 0x010101ff00000001, + 0x010101ff00000100, 0x010101ff01ffffff, 0x010101ff01ffff01, 0x010101ff01ff01ff, + 0x010101ff01ff0101, 0x010101ff01000000, 0x010101ff0101ffff, 0x010101ff0101ff01, + 0x010101ff010101ff, 0x010101ff01010101, 0x01010100ffff0000, 0x01010100ff0000ff, + 0x01010100ff000100, 0x01010100ff01ff00, 0x01010100ff010000, 0x0101010000ffff00, + 0x010101000000ffff, 0x0101010000000000, 0x0101010000000101, 0x010101000001ff00, + 0x0101010000010001, 0x0101010000010100, 0x010101000100ffff, 0x0101010001000001, + 0x01010101ffffffff, 0x01010101ffffff01, 0x01010101ffff01ff, 0x01010101ffff0101, + 0x01010101ff01ffff, 0x01010101ff01ff01, 0x01010101ff0101ff, 0x01010101ff010101, + 0x010101010000ff00, 0x01010101000000ff, 0x0101010100000001, 0x0101010101ffffff, + 0x0101010101ffff01, 0x0101010101ff01ff, 0x0101010101ff0101, 0x0101010101000000, + 0x010101010101ffff, 0x010101010101ff01, 0x01010101010101ff, 0x0101010101010101, +GGML_TABLE_END() +#else +GGML_TABLE_BEGIN(uint32_t, iq1s_grid_gpu, NGRID_IQ1S) + 0x00000000, 0x00000002, 0x00000101, 0x00000200, 0x00000202, 0x00010001, 0x00010101, 0x00020000, + 0x00020002, 0x00020200, 0x00020202, 0x01000101, 0x01010001, 0x01010100, 0x01010102, 0x01020101, + 0x02000000, 0x02000002, 0x02000200, 0x02000202, 0x02010101, 0x02020000, 0x02020002, 0x02020200, + 0x02020202, 0x00000110, 0x00000111, 0x00010011, 0x00010110, 0x00010112, 0x00010211, 0x00010212, + 0x00020111, 0x01000011, 0x01000112, 0x01000211, 0x01010012, 0x01010111, 0x01010212, 0x01020011, + 0x01020110, 0x01020112, 0x01020210, 0x02000111, 0x02010011, 0x02010110, 0x02010112, 0x02020111, + 0x00000020, 0x00000022, 0x00000220, 0x00000222, 0x00010121, 0x00020020, 0x00020022, 0x00020220, + 0x00020222, 0x01000121, 0x01010021, 0x01010221, 0x01020120, 0x01020221, 0x02000020, 0x02000022, + 0x02000220, 0x02000222, 0x02010021, 0x02010121, 0x02010221, 0x02020020, 0x02020022, 0x02020220, + 0x02020222, 0x00011001, 0x00011100, 0x00011102, 0x00021101, 0x01001001, 0x01001201, 0x01011101, + 0x01011202, 0x01021100, 0x01021101, 0x02011001, 0x02011201, 0x02021101, 0x00001011, 0x00001110, + 0x00001111, 0x00001112, 0x00011111, 0x00011210, 0x00011212, 0x00021211, 0x01001010, 0x01001111, + 0x01001212, 0x01011010, 0x01011011, 0x01011110, 0x01011111, 0x01011112, 0x01011211, 0x01021010, + 0x01021012, 0x01021111, 0x01021210, 0x01021212, 0x02001011, 0x02011011, 0x02011111, 0x02011210, + 0x02011212, 0x02021011, 0x02021110, 0x02021111, 0x02021112, 0x02021211, 0x00011120, 0x00011221, + 0x01001021, 0x01001120, 0x01011020, 0x01011022, 0x01011121, 0x01011220, 0x01021020, 0x01021021, + 0x01021122, 0x01021221, 0x02001121, 0x02011021, 0x02011120, 0x02011221, 0x00002000, 0x00002002, + 0x00002200, 0x00002202, 0x00012101, 0x00022000, 0x00022002, 0x00022200, 0x00022202, 0x01002101, + 0x01012001, 0x01012102, 0x01022101, 0x02002000, 0x02002002, 0x02002200, 0x02002202, 0x02012101, + 0x02022000, 0x02022002, 0x02022200, 0x02022202, 0x00002111, 0x00012011, 0x00012110, 0x00012211, + 0x00022110, 0x00022111, 0x01002011, 0x01012010, 0x01012011, 0x01012111, 0x01022011, 0x01022110, + 0x01022211, 0x02012011, 0x02012110, 0x02012112, 0x02012211, 0x02022111, 0x00002020, 0x00002022, + 0x00002220, 0x00002222, 0x00012121, 0x00022020, 0x00022022, 0x00022220, 0x00022222, 0x01002121, + 0x01012021, 0x01012221, 0x01022021, 0x01022121, 0x02002020, 0x02002022, 0x02002121, 0x02002220, + 0x02002222, 0x02012121, 0x02022020, 0x02022022, 0x02022220, 0x02022222, 0x00110000, 0x00110001, + 0x00110100, 0x00110201, 0x00120100, 0x00120101, 0x01100001, 0x01100100, 0x01110000, 0x01110101, + 0x01110200, 0x01120001, 0x01120100, 0x01120101, 0x01120201, 0x02110001, 0x02110100, 0x02110102, + 0x02120001, 0x02120101, 0x00100011, 0x00100110, 0x00100112, 0x00100211, 0x00110010, 0x00110012, + 0x00110111, 0x00110210, 0x00120011, 0x00120110, 0x00120211, 0x01100111, 0x01100212, 0x01110010, + 0x01110011, 0x01110012, 0x01110110, 0x01110111, 0x01110112, 0x01110211, 0x01120010, 0x01120111, + 0x02100110, 0x02110012, 0x02110111, 0x02120011, 0x02120110, 0x00110021, 0x00110120, 0x00110122, + 0x00120121, 0x01100020, 0x01100122, 0x01100221, 0x01110022, 0x01110121, 0x01110220, 0x01110222, + 0x01120120, 0x01120122, 0x02100121, 0x02110021, 0x02110120, 0x02110122, 0x02120121, 0x00101001, + 0x00101102, 0x00101201, 0x00111100, 0x00111101, 0x00111200, 0x00111201, 0x00121001, 0x00121102, + 0x01101001, 0x01101101, 0x01101102, 0x01101200, 0x01101202, 0x01111001, 0x01111100, 0x01111101, + 0x01111102, 0x01111201, 0x01121002, 0x01121101, 0x01121200, 0x02101100, 0x02101201, 0x02111000, + 0x02111100, 0x02111101, 0x02111200, 0x02111201, 0x02111202, 0x02121001, 0x02121100, 0x02121101, + 0x02121201, 0x00101012, 0x00101111, 0x00101212, 0x00111011, 0x00111110, 0x00111111, 0x00111112, + 0x00111211, 0x00121010, 0x00121012, 0x00121111, 0x00121210, 0x00121212, 0x01101011, 0x01101110, + 0x01101111, 0x01101112, 0x01111011, 0x01111012, 0x01111110, 0x01111111, 0x01111112, 0x01111211, + 0x01111212, 0x01121011, 0x01121110, 0x01121111, 0x01121112, 0x01121211, 0x02101010, 0x02101012, + 0x02101110, 0x02101111, 0x02101210, 0x02101212, 0x02111010, 0x02111011, 0x02111110, 0x02111111, + 0x02111112, 0x02111211, 0x02111212, 0x02121010, 0x02121012, 0x02121111, 0x00101021, 0x00101120, + 0x00101121, 0x00101122, 0x00111121, 0x00111122, 0x00111220, 0x00111222, 0x00121021, 0x00121122, + 0x01101020, 0x01101022, 0x01101120, 0x01101121, 0x01101220, 0x01101222, 0x01111021, 0x01111121, + 0x01111122, 0x01111220, 0x01111221, 0x01121021, 0x01121120, 0x01121121, 0x01121220, 0x01121221, + 0x01121222, 0x02101122, 0x02101222, 0x02111022, 0x02111121, 0x02121120, 0x02121221, 0x00112001, + 0x00112102, 0x00122101, 0x01102001, 0x01102100, 0x01102102, 0x01102201, 0x01112000, 0x01112101, + 0x01112200, 0x01112202, 0x01122000, 0x01122001, 0x01122100, 0x01122102, 0x01122201, 0x02102101, + 0x02112001, 0x02112100, 0x02122101, 0x00112010, 0x00112012, 0x00112111, 0x00112212, 0x00122011, + 0x00122111, 0x01102012, 0x01102110, 0x01102111, 0x01102210, 0x01112011, 0x01112110, 0x01112111, + 0x01112112, 0x01112211, 0x01112212, 0x01122010, 0x01122111, 0x01122212, 0x02102211, 0x02112011, + 0x02112012, 0x02112111, 0x02112210, 0x02122011, 0x02122112, 0x02122211, 0x00102221, 0x00112122, + 0x00122120, 0x00122122, 0x01102120, 0x01102122, 0x01102221, 0x01112020, 0x01112022, 0x01112121, + 0x01112220, 0x01122021, 0x01122122, 0x01122221, 0x02102121, 0x02112021, 0x02112122, 0x02112222, + 0x00200000, 0x00200002, 0x00200200, 0x00200202, 0x00210101, 0x00220000, 0x00220002, 0x00220101, + 0x00220200, 0x00220202, 0x01200101, 0x01210001, 0x01210201, 0x01220001, 0x01220101, 0x02200000, + 0x02200002, 0x02200200, 0x02200202, 0x02210101, 0x02220000, 0x02220002, 0x02220101, 0x02220200, + 0x02220202, 0x00200111, 0x00210011, 0x00210110, 0x00210211, 0x00220111, 0x01200012, 0x01200110, + 0x01200211, 0x01210111, 0x01210210, 0x01210212, 0x01220011, 0x01220110, 0x01220111, 0x01220112, + 0x02200111, 0x02210010, 0x02210112, 0x02210211, 0x02220111, 0x00200021, 0x00200220, 0x00200222, + 0x00210021, 0x00210121, 0x00220020, 0x00220022, 0x00220220, 0x00220222, 0x01200121, 0x01210021, + 0x01210122, 0x01210221, 0x01220121, 0x02200021, 0x02200220, 0x02200222, 0x02210021, 0x02210121, + 0x02220020, 0x02220022, 0x02220220, 0x02220222, 0x00201101, 0x00211100, 0x00211102, 0x00211201, + 0x00221101, 0x01201100, 0x01201101, 0x01201102, 0x01201201, 0x01211002, 0x01211101, 0x01211200, + 0x01211202, 0x01221102, 0x02201101, 0x02211001, 0x02211100, 0x02211201, 0x02221001, 0x02221101, + 0x00201211, 0x00211111, 0x00221011, 0x00221211, 0x01201010, 0x01201111, 0x01201210, 0x01211011, + 0x01211110, 0x01211111, 0x01211211, 0x01221012, 0x01221111, 0x01221210, 0x02201211, 0x02211010, + 0x02211110, 0x02211111, 0x02211210, 0x02211212, 0x02221011, 0x02221110, 0x02221112, 0x02221211, + 0x00201121, 0x00211020, 0x00211022, 0x00211221, 0x00221121, 0x01201021, 0x01201221, 0x01211121, + 0x01221020, 0x01221021, 0x01221221, 0x02201120, 0x02201122, 0x02211020, 0x02211222, 0x00202000, + 0x00202002, 0x00202200, 0x00202202, 0x00212101, 0x00222000, 0x00222002, 0x00222200, 0x00222202, + 0x01202101, 0x01212001, 0x01212100, 0x01222101, 0x02202000, 0x02202002, 0x02202200, 0x02202202, + 0x02222000, 0x02222002, 0x02222200, 0x02222202, 0x00202211, 0x00212011, 0x00212110, 0x00212211, + 0x00222111, 0x01202112, 0x01202211, 0x01212012, 0x01212111, 0x01222011, 0x01222110, 0x01222112, + 0x01222211, 0x02202111, 0x02212010, 0x02212112, 0x02212211, 0x02222110, 0x02222111, 0x00202020, + 0x00202022, 0x00202220, 0x00202222, 0x00222020, 0x00222022, 0x00222220, 0x00222222, 0x01202121, + 0x01212021, 0x01212122, 0x01212221, 0x01222121, 0x02202020, 0x02202022, 0x02202220, 0x02202222, + 0x02212121, 0x02222020, 0x02222022, 0x02222220, 0x02222222, 0x10000101, 0x10010001, 0x10010102, + 0x10020101, 0x11000201, 0x11010002, 0x11010101, 0x11010200, 0x11010202, 0x11020001, 0x11020100, + 0x11020102, 0x12010100, 0x12010201, 0x12020001, 0x12020102, 0x10000010, 0x10000011, 0x10000110, + 0x10000112, 0x10000211, 0x10010012, 0x10010111, 0x10010112, 0x10010210, 0x10010212, 0x10020011, + 0x10020112, 0x10020211, 0x11000111, 0x11000210, 0x11000212, 0x11010011, 0x11010110, 0x11010111, + 0x11010112, 0x11010211, 0x11010212, 0x11020111, 0x11020210, 0x11020212, 0x12000011, 0x12000110, + 0x12000112, 0x12010010, 0x12010012, 0x12010111, 0x12020010, 0x12020011, 0x12020012, 0x10000121, + 0x10010021, 0x10010120, 0x10010122, 0x10020121, 0x11000021, 0x11010022, 0x11010121, 0x11010222, + 0x11020120, 0x11020221, 0x12000221, 0x12010120, 0x12020121, 0x10001001, 0x10011101, 0x10011201, + 0x10021201, 0x11001101, 0x11001200, 0x11001202, 0x11011001, 0x11011100, 0x11011101, 0x11011102, + 0x11021001, 0x11021002, 0x11021101, 0x11021200, 0x11021202, 0x12001001, 0x12001102, 0x12001201, + 0x12011000, 0x12011002, 0x12011101, 0x12021000, 0x12021001, 0x12021201, 0x10001011, 0x10001012, + 0x10001111, 0x10001212, 0x10011011, 0x10011110, 0x10011111, 0x10011112, 0x10011211, 0x10021010, + 0x10021111, 0x10021212, 0x11001011, 0x11001110, 0x11001111, 0x11001112, 0x11001211, 0x11011010, + 0x11011011, 0x11011110, 0x11011111, 0x11011112, 0x11011210, 0x11011211, 0x11021011, 0x11021110, + 0x11021111, 0x11021112, 0x11021211, 0x12001012, 0x12001110, 0x12001111, 0x12001210, 0x12011011, + 0x12011110, 0x12011111, 0x12011112, 0x12011211, 0x12011212, 0x12021111, 0x12021210, 0x12021212, + 0x10001021, 0x10001121, 0x10001221, 0x10011120, 0x10011121, 0x10011220, 0x10011222, 0x10021021, + 0x10021120, 0x10021221, 0x11001020, 0x11001022, 0x11001121, 0x11001220, 0x11011020, 0x11011021, + 0x11011022, 0x11011121, 0x11011122, 0x11011221, 0x11021022, 0x11021121, 0x11021220, 0x12001021, + 0x12001121, 0x12001222, 0x12011120, 0x12011121, 0x12021021, 0x12021120, 0x12021122, 0x10002101, + 0x10012001, 0x10012101, 0x10012202, 0x10022101, 0x11002002, 0x11002201, 0x11012000, 0x11012101, + 0x11012200, 0x11022001, 0x11022100, 0x11022102, 0x11022201, 0x12002101, 0x12012001, 0x12012100, + 0x12012102, 0x12012201, 0x12022101, 0x10002011, 0x10002111, 0x10002112, 0x10002212, 0x10012010, + 0x10012110, 0x10012111, 0x10012210, 0x10022011, 0x10022110, 0x10022112, 0x11002010, 0x11002111, + 0x11002212, 0x11012011, 0x11012012, 0x11012110, 0x11012111, 0x11012112, 0x11012211, 0x11022010, + 0x11022012, 0x11022111, 0x11022112, 0x11022212, 0x12002112, 0x12002211, 0x12012012, 0x12012111, + 0x12012112, 0x12012210, 0x12022011, 0x12022110, 0x12022112, 0x12022211, 0x10012122, 0x11002120, + 0x11002122, 0x11002221, 0x11012121, 0x11012220, 0x11012222, 0x11022120, 0x11022221, 0x12012120, + 0x12022121, 0x10100001, 0x10100100, 0x10100101, 0x10100102, 0x10100201, 0x10110002, 0x10110101, + 0x10110202, 0x10120001, 0x10120100, 0x10120201, 0x11100000, 0x11100101, 0x11100200, 0x11110001, + 0x11110100, 0x11110101, 0x11110102, 0x11110201, 0x11120101, 0x11120200, 0x12100102, 0x12100201, + 0x12110101, 0x12110200, 0x12120000, 0x12120001, 0x12120102, 0x12120201, 0x10100111, 0x10100210, + 0x10100211, 0x10100212, 0x10110011, 0x10110110, 0x10110111, 0x10110112, 0x10110210, 0x10110211, + 0x10120010, 0x10120111, 0x10120112, 0x10120210, 0x10120212, 0x11100011, 0x11100110, 0x11100111, + 0x11100112, 0x11100211, 0x11110010, 0x11110011, 0x11110012, 0x11110110, 0x11110111, 0x11110112, + 0x11110210, 0x11110211, 0x11110212, 0x11120011, 0x11120110, 0x11120111, 0x11120112, 0x11120211, + 0x12100012, 0x12100111, 0x12110011, 0x12110110, 0x12110111, 0x12110112, 0x12110211, 0x12120010, + 0x12120111, 0x12120212, 0x10100021, 0x10100122, 0x10110022, 0x10110121, 0x10110222, 0x10120021, + 0x10120120, 0x11100022, 0x11100121, 0x11100222, 0x11110021, 0x11110120, 0x11110121, 0x11110122, + 0x11110221, 0x11120022, 0x11120121, 0x12100121, 0x12110020, 0x12110022, 0x12110121, 0x12110221, + 0x12110222, 0x12120120, 0x10101100, 0x10101101, 0x10111001, 0x10111100, 0x10111101, 0x10111102, + 0x10111200, 0x10111201, 0x10121001, 0x10121101, 0x10121200, 0x10121202, 0x11101001, 0x11101100, + 0x11101101, 0x11101102, 0x11101201, 0x11101202, 0x11111000, 0x11111001, 0x11111100, 0x11111101, + 0x11111102, 0x11111200, 0x11111201, 0x11111202, 0x11121001, 0x11121002, 0x11121100, 0x11121101, + 0x11121102, 0x11121201, 0x12101000, 0x12101200, 0x12101202, 0x12111001, 0x12111100, 0x12111101, + 0x12111102, 0x12111201, 0x12121001, 0x12121100, 0x12121101, 0x12121202, 0x10101011, 0x10101012, + 0x10101110, 0x10101111, 0x10101112, 0x10101211, 0x10111010, 0x10111011, 0x10111012, 0x10111110, + 0x10111111, 0x10111112, 0x10111211, 0x10111212, 0x10121011, 0x10121110, 0x10121111, 0x10121112, + 0x10121211, 0x11101010, 0x11101011, 0x11101012, 0x11101110, 0x11101111, 0x11101112, 0x11101210, + 0x11101211, 0x11111010, 0x11111011, 0x11111012, 0x11111110, 0x11111111, 0x11111112, 0x11111210, + 0x11111211, 0x11111212, 0x11121010, 0x11121011, 0x11121110, 0x11121111, 0x11121112, 0x11121210, + 0x11121211, 0x11121212, 0x12101011, 0x12101110, 0x12101111, 0x12101211, 0x12101212, 0x12111010, + 0x12111011, 0x12111110, 0x12111111, 0x12111112, 0x12111210, 0x12111211, 0x12121011, 0x12121110, + 0x12121111, 0x12121112, 0x12121211, 0x10101020, 0x10101021, 0x10101022, 0x10101120, 0x10101122, + 0x10101220, 0x10101221, 0x10111021, 0x10111120, 0x10111121, 0x10111220, 0x10111221, 0x10121020, + 0x10121021, 0x10121022, 0x10121120, 0x10121121, 0x10121122, 0x10121220, 0x10121221, 0x11101021, + 0x11101121, 0x11101122, 0x11101220, 0x11101221, 0x11101222, 0x11111020, 0x11111021, 0x11111022, + 0x11111120, 0x11111121, 0x11111122, 0x11111220, 0x11111221, 0x11111222, 0x11121021, 0x11121120, + 0x11121121, 0x11121221, 0x12101022, 0x12101121, 0x12101122, 0x12101220, 0x12101221, 0x12101222, + 0x12111021, 0x12111121, 0x12111222, 0x12121022, 0x12121121, 0x12121122, 0x12121220, 0x12121221, + 0x10102100, 0x10102101, 0x10102102, 0x10102201, 0x10112000, 0x10112101, 0x10112200, 0x10122001, + 0x10122202, 0x11102101, 0x11102200, 0x11102202, 0x11112001, 0x11112100, 0x11112101, 0x11112102, + 0x11112200, 0x11112201, 0x11122000, 0x11122002, 0x11122100, 0x11122101, 0x12102002, 0x12102201, + 0x12112000, 0x12112002, 0x12112101, 0x12112200, 0x12122001, 0x12122201, 0x10102011, 0x10102012, + 0x10102111, 0x10102212, 0x10112011, 0x10112110, 0x10112111, 0x10112112, 0x10112211, 0x10122111, + 0x11102011, 0x11102110, 0x11102111, 0x11102112, 0x11102211, 0x11112010, 0x11112011, 0x11112012, + 0x11112110, 0x11112111, 0x11112112, 0x11112210, 0x11112211, 0x11112212, 0x11122011, 0x11122110, + 0x11122111, 0x11122112, 0x11122211, 0x12102011, 0x12102111, 0x12102211, 0x12112011, 0x12112110, + 0x12112111, 0x12112112, 0x12112210, 0x12112211, 0x12122111, 0x10102120, 0x10102220, 0x10112121, + 0x10112222, 0x10122020, 0x10122121, 0x10122122, 0x10122221, 0x11102121, 0x11102220, 0x11102221, + 0x11112021, 0x11112121, 0x11112122, 0x11112220, 0x11112221, 0x11122022, 0x11122121, 0x11122220, + 0x11122222, 0x12102021, 0x12102222, 0x12112022, 0x12112121, 0x12112122, 0x12112220, 0x12112222, + 0x12122021, 0x10200101, 0x10210100, 0x10210102, 0x10210201, 0x10220101, 0x11200100, 0x11210000, + 0x11210101, 0x11210102, 0x11210200, 0x11210202, 0x11220001, 0x11220100, 0x11220102, 0x11220201, + 0x12200001, 0x12210102, 0x12220101, 0x10200011, 0x10200110, 0x10200112, 0x10200211, 0x10210012, + 0x10210111, 0x10220011, 0x10220012, 0x10220112, 0x10220211, 0x11200111, 0x11200211, 0x11210011, + 0x11210111, 0x11210112, 0x11210211, 0x11220111, 0x11220112, 0x11220212, 0x12200110, 0x12200212, + 0x12210012, 0x12210111, 0x12220011, 0x12220112, 0x12220211, 0x10210021, 0x10210122, 0x10210221, + 0x11200020, 0x11200021, 0x11200122, 0x11210121, 0x11210122, 0x11210220, 0x11220020, 0x12200121, + 0x12210021, 0x12210122, 0x12220121, 0x10211001, 0x10211002, 0x10211101, 0x10211102, 0x10211202, + 0x10221001, 0x10221102, 0x10221201, 0x11201000, 0x11201002, 0x11201101, 0x11201200, 0x11201202, + 0x11211001, 0x11211100, 0x11211101, 0x11211102, 0x11211201, 0x11211202, 0x11221000, 0x11221002, + 0x11221101, 0x12201100, 0x12201101, 0x12201201, 0x12211000, 0x12211002, 0x12211100, 0x12211101, + 0x12211102, 0x12211200, 0x12211202, 0x12221001, 0x12221100, 0x12221201, 0x10201111, 0x10201210, + 0x10201212, 0x10211011, 0x10211111, 0x10211112, 0x10211211, 0x11201110, 0x11201111, 0x11201112, + 0x11201211, 0x11211010, 0x11211011, 0x11211110, 0x11211111, 0x11211112, 0x11211211, 0x11221011, + 0x11221110, 0x11221111, 0x11221112, 0x11221211, 0x12201112, 0x12201211, 0x12201212, 0x12211011, + 0x12211111, 0x12211112, 0x12211211, 0x12211212, 0x12221012, 0x12221111, 0x12221112, 0x12221210, + 0x10201022, 0x10201221, 0x10211121, 0x10221020, 0x10221122, 0x10221220, 0x10221221, 0x11201020, + 0x11201121, 0x11201220, 0x11201222, 0x11211021, 0x11211120, 0x11211121, 0x11211122, 0x11211220, + 0x11211222, 0x11221020, 0x11221121, 0x11221220, 0x12201020, 0x12201022, 0x12201121, 0x12201222, + 0x12211120, 0x12211122, 0x12211220, 0x12211221, 0x12221020, 0x12221120, 0x12221122, 0x12221222, + 0x10212102, 0x10212201, 0x10222101, 0x11202001, 0x11212002, 0x11212101, 0x11212202, 0x11222001, + 0x11222201, 0x12202101, 0x12212001, 0x12212200, 0x12222102, 0x10202011, 0x10202110, 0x10212010, + 0x10212111, 0x10222011, 0x10222110, 0x10222112, 0x10222211, 0x11202010, 0x11202011, 0x11202111, + 0x11202112, 0x11202210, 0x11212011, 0x11212110, 0x11212111, 0x11212112, 0x11212211, 0x11222010, + 0x11222111, 0x11222212, 0x12202012, 0x12202110, 0x12202212, 0x12212111, 0x12222011, 0x12222110, + 0x12222111, 0x12222211, 0x10212021, 0x10212122, 0x10212220, 0x11202021, 0x11202120, 0x11202221, + 0x11212020, 0x11212121, 0x11212220, 0x11212222, 0x11222120, 0x11222121, 0x11222221, 0x12202122, + 0x12212120, 0x12212220, 0x12212222, 0x12222122, 0x20000000, 0x20000002, 0x20000200, 0x20000202, + 0x20020000, 0x20020002, 0x20020200, 0x20020202, 0x21000101, 0x21010000, 0x21010001, 0x21010100, + 0x21010102, 0x21010201, 0x21020101, 0x22000000, 0x22000002, 0x22000200, 0x22000202, 0x22010101, + 0x22020000, 0x22020002, 0x22020200, 0x22020202, 0x20000111, 0x20010011, 0x20010110, 0x20010112, + 0x20010211, 0x20020111, 0x21000011, 0x21000110, 0x21000211, 0x21010010, 0x21010012, 0x21010111, + 0x21010112, 0x21010210, 0x21010211, 0x21020110, 0x21020112, 0x21020211, 0x22000111, 0x22000211, + 0x22010110, 0x22010112, 0x22010211, 0x22020111, 0x20000020, 0x20000022, 0x20000220, 0x20000222, + 0x20010121, 0x20020020, 0x20020022, 0x20020220, 0x20020222, 0x21010021, 0x21010120, 0x21010221, + 0x21020121, 0x22000020, 0x22000022, 0x22000220, 0x22000222, 0x22010121, 0x22020020, 0x22020022, + 0x22020220, 0x22020222, 0x20011100, 0x20011201, 0x21001001, 0x21001100, 0x21011001, 0x21011101, + 0x21011202, 0x21021001, 0x21021100, 0x21021201, 0x22011100, 0x22011201, 0x20001011, 0x20001211, + 0x20011012, 0x20011111, 0x20011212, 0x20021112, 0x20021211, 0x21001010, 0x21001011, 0x21001111, + 0x21001210, 0x21011011, 0x21011110, 0x21011111, 0x21011112, 0x21011211, 0x21011212, 0x21021111, + 0x21021112, 0x21021210, 0x21021212, 0x22001011, 0x22001110, 0x22001112, 0x22001211, 0x22011010, + 0x22011012, 0x22011111, 0x22011210, 0x22021112, 0x20011021, 0x20011122, 0x20011221, 0x20021121, + 0x21001021, 0x21001120, 0x21001221, 0x21001222, 0x21011020, 0x21011121, 0x21011221, 0x21011222, + 0x21021021, 0x21021122, 0x21021222, 0x22001121, 0x22011021, 0x22011222, 0x22021120, 0x20002000, + 0x20002002, 0x20002200, 0x20002202, 0x20012101, 0x20022000, 0x20022002, 0x20022200, 0x20022202, + 0x21002001, 0x21002101, 0x21012001, 0x21012100, 0x21012201, 0x21022101, 0x21022201, 0x22002000, + 0x22002002, 0x22002200, 0x22002202, 0x22012101, 0x22022000, 0x22022002, 0x22022200, 0x22022202, + 0x20002111, 0x20002112, 0x20012011, 0x20012110, 0x20012112, 0x20022111, 0x21002011, 0x21002110, + 0x21002112, 0x21002211, 0x21012010, 0x21012012, 0x21012111, 0x21012212, 0x21022011, 0x21022110, + 0x22002111, 0x22012112, 0x22012211, 0x22022111, 0x20002020, 0x20002022, 0x20002220, 0x20002222, + 0x20012121, 0x20022020, 0x20022022, 0x20022220, 0x20022222, 0x21002121, 0x21012021, 0x21012120, + 0x21012122, 0x22002020, 0x22002022, 0x22002220, 0x22002222, 0x22012121, 0x22022020, 0x22022022, + 0x22022220, 0x22022222, 0x20100101, 0x20110001, 0x20110102, 0x20110200, 0x20110201, 0x20120101, + 0x21100001, 0x21100102, 0x21100201, 0x21110101, 0x21110200, 0x21110202, 0x21120201, 0x21120202, + 0x22100101, 0x22110001, 0x22110100, 0x22110102, 0x22110201, 0x22120101, 0x20100011, 0x20100110, + 0x20100112, 0x20100211, 0x20110010, 0x20110111, 0x20110210, 0x20110212, 0x20120011, 0x20120110, + 0x20120112, 0x20120211, 0x21100010, 0x21100111, 0x21110010, 0x21110011, 0x21110110, 0x21110111, + 0x21110112, 0x21110211, 0x21120012, 0x21120111, 0x22100110, 0x22100112, 0x22110012, 0x22110111, + 0x22110210, 0x22120011, 0x22120110, 0x22120112, 0x22120211, 0x20100121, 0x20110021, 0x20110120, + 0x20110221, 0x20120121, 0x21100120, 0x21100122, 0x21100221, 0x21110020, 0x21110022, 0x21110121, + 0x21110220, 0x21120122, 0x21120221, 0x22100121, 0x22110120, 0x22110122, 0x22120221, 0x20101001, + 0x20101100, 0x20101102, 0x20111000, 0x20111101, 0x20111200, 0x20121102, 0x21101000, 0x21101202, + 0x21111001, 0x21111100, 0x21111101, 0x21111102, 0x21111200, 0x21111201, 0x21121000, 0x21121001, + 0x21121002, 0x21121101, 0x22101100, 0x22101102, 0x22111002, 0x22111100, 0x22111101, 0x22111200, + 0x22121001, 0x22121201, 0x20101010, 0x20101111, 0x20101210, 0x20101212, 0x20111010, 0x20111011, + 0x20111110, 0x20111111, 0x20111112, 0x20111211, 0x20121011, 0x20121111, 0x20121211, 0x20121212, + 0x21101011, 0x21101110, 0x21101111, 0x21101112, 0x21101211, 0x21111010, 0x21111011, 0x21111012, + 0x21111110, 0x21111111, 0x21111112, 0x21111210, 0x21111211, 0x21111212, 0x21121011, 0x21121110, + 0x21121111, 0x21121112, 0x21121211, 0x22101011, 0x22101111, 0x22101210, 0x22111011, 0x22111012, + 0x22111110, 0x22111111, 0x22111112, 0x22111211, 0x22111212, 0x22121010, 0x22121012, 0x22121111, + 0x22121210, 0x22121212, 0x20101021, 0x20101120, 0x20111020, 0x20111121, 0x20111221, 0x20121020, + 0x20121122, 0x20121221, 0x21101121, 0x21101220, 0x21101221, 0x21111021, 0x21111022, 0x21111121, + 0x21111122, 0x21111221, 0x21121121, 0x21121220, 0x22101022, 0x22101120, 0x22101221, 0x22101222, + 0x22111022, 0x22111120, 0x22111121, 0x22121120, 0x22121122, 0x22121221, 0x20102101, 0x20112102, + 0x20112201, 0x20122101, 0x21102001, 0x21102102, 0x21112000, 0x21112002, 0x21112101, 0x21112102, + 0x21112202, 0x21122100, 0x21122101, 0x22102101, 0x22112001, 0x22112102, 0x22112201, 0x22122101, + 0x20102110, 0x20102112, 0x20102211, 0x20112010, 0x20112012, 0x20112111, 0x20112210, 0x20112212, + 0x20122010, 0x20122011, 0x20122110, 0x20122112, 0x21102010, 0x21102012, 0x21102111, 0x21102210, + 0x21102212, 0x21112011, 0x21112110, 0x21112111, 0x21112112, 0x21112211, 0x21122012, 0x21122111, + 0x21122112, 0x21122212, 0x22102011, 0x22102110, 0x22112010, 0x22112012, 0x22112111, 0x22112212, + 0x22122011, 0x22122112, 0x20102121, 0x20112121, 0x20122121, 0x21102120, 0x21102122, 0x21102221, + 0x21112020, 0x21112121, 0x21112220, 0x21122021, 0x22102121, 0x22112021, 0x22112120, 0x22112121, + 0x22112122, 0x20200000, 0x20200002, 0x20200200, 0x20200202, 0x20210101, 0x20220000, 0x20220002, + 0x20220200, 0x20220202, 0x21200101, 0x21210001, 0x21210100, 0x21210102, 0x21210201, 0x22200000, + 0x22200002, 0x22200200, 0x22200202, 0x22210101, 0x22220000, 0x22220002, 0x22220200, 0x22220202, + 0x20200111, 0x20200211, 0x20210011, 0x20210110, 0x20210112, 0x20210211, 0x20210212, 0x21200112, + 0x21200211, 0x21210011, 0x21210111, 0x21210210, 0x21210212, 0x21220011, 0x21220110, 0x22200111, + 0x22210010, 0x22210012, 0x22210112, 0x22210211, 0x20200022, 0x20200220, 0x20200222, 0x20210020, + 0x20210221, 0x20220022, 0x20220220, 0x20220222, 0x21200121, 0x21210021, 0x21210122, 0x21210221, + 0x21220121, 0x22200020, 0x22200022, 0x22200220, 0x22200222, 0x22210121, 0x22220020, 0x22220022, + 0x22220220, 0x22220222, 0x20211201, 0x20221101, 0x21201001, 0x21201100, 0x21211000, 0x21211100, + 0x21211101, 0x21211200, 0x21211202, 0x21221001, 0x21221101, 0x21221102, 0x21221200, 0x21221201, + 0x22201101, 0x20201112, 0x20201211, 0x20211010, 0x20211012, 0x20211111, 0x20211210, 0x20221112, + 0x20221211, 0x21201012, 0x21201111, 0x21211011, 0x21211110, 0x21211111, 0x21211112, 0x21211211, + 0x21221111, 0x21221212, 0x22201011, 0x22201110, 0x22201111, 0x22201112, 0x22201211, 0x22211012, + 0x22211111, 0x22211210, 0x20201121, 0x20211021, 0x20211122, 0x20211222, 0x20221021, 0x20221121, + 0x21201120, 0x21201122, 0x21201222, 0x21211022, 0x21211121, 0x21211122, 0x21211220, 0x21221020, + 0x21221022, 0x22201122, 0x22211020, 0x22211121, 0x22211122, 0x22211221, 0x22221021, 0x22221120, + 0x22221122, 0x20202000, 0x20202002, 0x20202200, 0x20202202, 0x20222000, 0x20222002, 0x20222200, + 0x20222202, 0x21212001, 0x21212100, 0x21212102, 0x21212201, 0x22202000, 0x22202002, 0x22202200, + 0x22202202, 0x22212101, 0x22222000, 0x22222002, 0x22222200, 0x22222202, 0x20202111, 0x20212110, + 0x20212211, 0x20222011, 0x20222111, 0x21202011, 0x21212010, 0x21212111, 0x21212212, 0x21222011, + 0x21222112, 0x21222211, 0x22212010, 0x22212112, 0x20202020, 0x20202022, 0x20202220, 0x20202222, + 0x20222020, 0x20222022, 0x20222220, 0x20222222, 0x21212021, 0x21212120, 0x21212122, 0x22202020, + 0x22202022, 0x22202220, 0x22202222, 0x22212121, 0x22222020, 0x22222022, 0x22222220, 0x22222222, +GGML_TABLE_END() +#endif + +#endif // GGML_COMMON_IMPL +#endif // GGML_COMMON_IMPL diff --git a/bindings/ruby/ext/ggml-cuda.h b/bindings/ruby/ext/ggml-cuda.h new file mode 100644 index 00000000000..5eb4af40f4d --- /dev/null +++ b/bindings/ruby/ext/ggml-cuda.h @@ -0,0 +1,43 @@ +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#ifdef GGML_USE_HIPBLAS +#define GGML_CUDA_NAME "ROCm" +#define GGML_CUBLAS_NAME "hipBLAS" +#else +#define GGML_CUDA_NAME "CUDA" +#define GGML_CUBLAS_NAME "cuBLAS" +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +#define GGML_CUDA_MAX_DEVICES 16 + +// backend API +GGML_API GGML_CALL ggml_backend_t ggml_backend_cuda_init(int device); + +GGML_API GGML_CALL bool ggml_backend_is_cuda(ggml_backend_t backend); + +// device buffer +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_cuda_buffer_type(int device); + +// split tensor buffer that splits matrices by rows across multiple devices +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_cuda_split_buffer_type(const float * tensor_split); + +// pinned host buffer for use with the CPU backend for faster copies between CPU and GPU +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_cuda_host_buffer_type(void); + +GGML_API GGML_CALL int ggml_backend_cuda_get_device_count(void); +GGML_API GGML_CALL void ggml_backend_cuda_get_device_description(int device, char * description, size_t description_size); +GGML_API GGML_CALL void ggml_backend_cuda_get_device_memory(int device, size_t * free, size_t * total); + +GGML_API GGML_CALL bool ggml_backend_cuda_register_host_buffer(void * buffer, size_t size); +GGML_API GGML_CALL void ggml_backend_cuda_unregister_host_buffer(void * buffer); + +#ifdef __cplusplus +} +#endif diff --git a/bindings/ruby/ext/ggml-impl.h b/bindings/ruby/ext/ggml-impl.h index d88f261449f..93a4f1a2b72 100644 --- a/bindings/ruby/ext/ggml-impl.h +++ b/bindings/ruby/ext/ggml-impl.h @@ -5,6 +5,7 @@ // GGML internal header #include +#include // load `stdlib.h` before other headers to work around MinGW bug: https://sourceforge.net/p/mingw-w64/bugs/192/ #include #include #include // memcpy @@ -18,6 +19,7 @@ extern "C" { // fall back to the _Static_assert C11 keyword. // if C99 - static_assert is noop // ref: https://stackoverflow.com/a/53923785/4039976 +#ifndef __cplusplus #ifndef static_assert #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201100L) #define static_assert(cond, msg) _Static_assert(cond, msg) @@ -25,6 +27,7 @@ extern "C" { #define static_assert(cond, msg) struct global_scope_noop_trick #endif #endif +#endif // __FMA__ and __F16C__ are not defined in MSVC, however they are implied with AVX2/AVX512 #if defined(_MSC_VER) && (defined(__AVX2__) || defined(__AVX512F__)) @@ -34,16 +37,17 @@ extern "C" { #ifndef __F16C__ #define __F16C__ #endif +#endif + +// __SSE3__ and __SSSE3__ are not defined in MSVC, but SSE3/SSSE3 are present when AVX/AVX2/AVX512 are available +#if defined(_MSC_VER) && (defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__)) #ifndef __SSE3__ #define __SSE3__ #endif +#ifndef __SSSE3__ +#define __SSSE3__ +#endif #endif - -#undef MIN -#undef MAX - -#define MIN(a, b) ((a) < (b) ? (a) : (b)) -#define MAX(a, b) ((a) > (b) ? (a) : (b)) // 16-bit float // on Arm, we use __fp16 @@ -56,14 +60,30 @@ extern "C" { // #include -#define GGML_COMPUTE_FP16_TO_FP32(x) ((float) (x)) -#define GGML_COMPUTE_FP32_TO_FP16(x) (x) +typedef __fp16 ggml_fp16_internal_t; -#define GGML_FP16_TO_FP32(x) ((float) (x)) -#define GGML_FP32_TO_FP16(x) (x) +#define GGML_COMPUTE_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) +#define GGML_COMPUTE_FP32_TO_FP16(x) ggml_compute_fp32_to_fp16(x) + +#define GGML_FP16_TO_FP32(x) ggml_compute_fp16_to_fp32(x) + +static inline float ggml_compute_fp16_to_fp32(ggml_fp16_t h) { + ggml_fp16_internal_t tmp; + memcpy(&tmp, &h, sizeof(ggml_fp16_t)); + return (float)tmp; +} + +static inline ggml_fp16_t ggml_compute_fp32_to_fp16(float f) { + ggml_fp16_t res; + ggml_fp16_internal_t tmp = f; + memcpy(&res, &tmp, sizeof(ggml_fp16_t)); + return res; +} #else +typedef uint16_t ggml_fp16_internal_t; + #ifdef __wasm_simd128__ #include #else @@ -217,8 +237,7 @@ extern float ggml_table_f32_f16[1 << 16]; // On ARM NEON, it's quicker to directly convert x -> x instead of calling into ggml_lookup_fp16_to_fp32, // so we define GGML_FP16_TO_FP32 and GGML_FP32_TO_FP16 elsewhere for NEON. // This is also true for POWER9. -#if !defined(GGML_FP16_TO_FP32) || !defined(GGML_FP32_TO_FP16) - +#if !defined(GGML_FP16_TO_FP32) inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) { uint16_t s; memcpy(&s, &f, sizeof(uint16_t)); @@ -226,19 +245,23 @@ inline static float ggml_lookup_fp16_to_fp32(ggml_fp16_t f) { } #define GGML_FP16_TO_FP32(x) ggml_lookup_fp16_to_fp32(x) -#define GGML_FP32_TO_FP16(x) GGML_COMPUTE_FP32_TO_FP16(x) +#endif +#if !defined(GGML_FP32_TO_FP16) +#define GGML_FP32_TO_FP16(x) GGML_COMPUTE_FP32_TO_FP16(x) #endif #define GGML_HASHTABLE_FULL ((size_t)-1) #define GGML_HASHTABLE_ALREADY_EXISTS ((size_t)-2) +struct ggml_hash_set ggml_hash_set_new(size_t size); + bool ggml_hash_contains (const struct ggml_hash_set hash_set, struct ggml_tensor * key); // returns GGML_HASHTABLE_FULL if table is full, otherwise the current index of the key or where it should be inserted size_t ggml_hash_find (const struct ggml_hash_set hash_set, struct ggml_tensor * key); -// returns GGML_HAHSHTABLE_ALREADY_EXISTS if key already exists, index otherwise, asserts if table is full +// returns GGML_HASHTABLE_ALREADY_EXISTS if key already exists, index otherwise, asserts if table is full size_t ggml_hash_insert ( struct ggml_hash_set hash_set, struct ggml_tensor * key); // return index, asserts if table is full diff --git a/bindings/ruby/ext/ggml-kompute.h b/bindings/ruby/ext/ggml-kompute.h new file mode 100644 index 00000000000..171465456a5 --- /dev/null +++ b/bindings/ruby/ext/ggml-kompute.h @@ -0,0 +1,46 @@ +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#include +#include +#include + +#ifdef __cplusplus +extern "C" { +#endif + +struct ggml_vk_device { + int index; + int type; // same as VkPhysicalDeviceType + size_t heapSize; + const char * name; + const char * vendor; + int subgroupSize; + uint64_t bufferAlignment; + uint64_t maxAlloc; +}; + +struct ggml_vk_device * ggml_vk_available_devices(size_t memoryRequired, size_t * count); +bool ggml_vk_get_device(struct ggml_vk_device * device, size_t memoryRequired, const char * name); +bool ggml_vk_has_vulkan(void); +bool ggml_vk_has_device(void); +struct ggml_vk_device ggml_vk_current_device(void); + +// +// backend API +// + +// forward declaration +typedef struct ggml_backend * ggml_backend_t; + +GGML_API ggml_backend_t ggml_backend_kompute_init(int device); + +GGML_API bool ggml_backend_is_kompute(ggml_backend_t backend); + +GGML_API ggml_backend_buffer_type_t ggml_backend_kompute_buffer_type(int device); + +#ifdef __cplusplus +} +#endif diff --git a/bindings/ruby/ext/ggml-metal.h b/bindings/ruby/ext/ggml-metal.h new file mode 100644 index 00000000000..a5c542189c2 --- /dev/null +++ b/bindings/ruby/ext/ggml-metal.h @@ -0,0 +1,66 @@ +// An interface allowing to compute ggml_cgraph with Metal +// +// This is a fully functional interface that extends ggml with GPU support for Apple devices. +// A similar interface can be created for other GPU backends (e.g. Vulkan, CUDA, OpenCL, etc.) +// +// How it works? +// +// As long as your program can create and evaluate a ggml_cgraph on the CPU, you can use this +// interface to evaluate the same graph on the GPU. Instead of using ggml_graph_compute(), you +// use ggml_metal_graph_compute() (or ggml_vulkan_graph_compute(), etc.) +// +// You only need to make sure that all memory buffers that you used during the graph creation +// are mapped to the device memory with the ggml_metal_add_buffer() function. This mapping is +// used during the graph evaluation to determine the arguments of the compute kernels. +// +// Synchronization between device and host memory (for example for input and output tensors) +// is done with the ggml_metal_set_tensor() and ggml_metal_get_tensor() functions. +// + +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#include +#include + +// max memory buffers that can be mapped to the device +#define GGML_METAL_MAX_BUFFERS 64 + +struct ggml_tensor; +struct ggml_cgraph; + +#ifdef __cplusplus +extern "C" { +#endif + +// +// backend API +// user-code should use only these functions +// + +GGML_API void ggml_backend_metal_log_set_callback(ggml_log_callback log_callback, void * user_data); + +GGML_API ggml_backend_t ggml_backend_metal_init(void); + +GGML_API bool ggml_backend_is_metal(ggml_backend_t backend); + +GGML_API GGML_CALL ggml_backend_buffer_t ggml_backend_metal_buffer_from_ptr(void * data, size_t size, size_t max_size); + +GGML_API void ggml_backend_metal_set_n_cb(ggml_backend_t backend, int n_cb); + +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_metal_buffer_type(void); + +// helper to check if the device supports a specific family +// ideally, the user code should be doing these checks +// ref: https://developer.apple.com/metal/Metal-Feature-Set-Tables.pdf +GGML_API bool ggml_backend_metal_supports_family(ggml_backend_t backend, int family); + +// capture all command buffers committed the next time `ggml_backend_graph_compute` is called +GGML_API void ggml_backend_metal_capture_next_compute(ggml_backend_t backend); + +#ifdef __cplusplus +} +#endif + diff --git a/bindings/ruby/ext/ggml-opencl.h b/bindings/ruby/ext/ggml-opencl.h new file mode 100644 index 00000000000..257a6be6af5 --- /dev/null +++ b/bindings/ruby/ext/ggml-opencl.h @@ -0,0 +1,36 @@ +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#ifdef __cplusplus +extern "C" { +#endif + +GGML_API void ggml_cl_init(void); + +GGML_API void ggml_cl_mul(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst); +GGML_API void ggml_cl_add(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst); +GGML_API bool ggml_cl_can_mul_mat(const struct ggml_tensor * src0, const struct ggml_tensor * src1, const struct ggml_tensor * dst); +GGML_API size_t ggml_cl_mul_mat_get_wsize(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst); +GGML_API void ggml_cl_mul_mat(const struct ggml_tensor * src0, const struct ggml_tensor * src1, struct ggml_tensor * dst, void * wdata, size_t wsize); + +// GGML_API void * ggml_cl_host_malloc(size_t size); +// GGML_API void ggml_cl_host_free(void * ptr); + +GGML_API void ggml_cl_free_data(const struct ggml_tensor* tensor); + +GGML_API void ggml_cl_transform_tensor(void * data, struct ggml_tensor * tensor); + +// backend API + +// GGML_API ggml_backend_t ggml_backend_opencl_init(void); + +// GGML_API bool ggml_backend_is_opencl(ggml_backend_t backend); + +GGML_API ggml_backend_buffer_type_t ggml_backend_opencl_buffer_type(void); +// GGML_API ggml_backend_buffer_type_t ggml_backend_opencl_host_buffer_type(void); + +#ifdef __cplusplus +} +#endif diff --git a/bindings/ruby/ext/ggml-quants.c b/bindings/ruby/ext/ggml-quants.c index 740be6dc5c7..32e84434a8c 100644 --- a/bindings/ruby/ext/ggml-quants.c +++ b/bindings/ruby/ext/ggml-quants.c @@ -1,10 +1,18 @@ +#define GGML_COMMON_IMPL_C +#include "ggml-common.h" + #include "ggml-quants.h" #include "ggml-impl.h" +#define GGML_COMMON_IMPL_C +#include "ggml-common.h" + #include #include #include #include +#include // for qsort +#include // for GGML_ASSERT #ifdef __ARM_NEON @@ -14,32 +22,12 @@ // #include -#if !defined(__aarch64__) -inline static int32_t vaddvq_s16(int16x8_t v) { - return - (int32_t)vgetq_lane_s16(v, 0) + (int32_t)vgetq_lane_s16(v, 1) + - (int32_t)vgetq_lane_s16(v, 2) + (int32_t)vgetq_lane_s16(v, 3) + - (int32_t)vgetq_lane_s16(v, 4) + (int32_t)vgetq_lane_s16(v, 5) + - (int32_t)vgetq_lane_s16(v, 6) + (int32_t)vgetq_lane_s16(v, 7); -} - -inline static int16x8_t vpaddq_s16(int16x8_t a, int16x8_t b) { - int16x4_t a0 = vpadd_s16(vget_low_s16(a), vget_high_s16(a)); - int16x4_t b0 = vpadd_s16(vget_low_s16(b), vget_high_s16(b)); - return vcombine_s16(a0, b0); -} - -inline static int32_t vaddvq_s32(int32x4_t v) { - return vgetq_lane_s32(v, 0) + vgetq_lane_s32(v, 1) + vgetq_lane_s32(v, 2) + vgetq_lane_s32(v, 3); -} -#endif - #else #ifdef __wasm_simd128__ #include #else -#ifdef __POWER9_VECTOR__ +#if defined(__POWER9_VECTOR__) || defined(__powerpc64__) #include #undef bool #define bool _Bool @@ -47,13 +35,15 @@ inline static int32_t vaddvq_s32(int32x4_t v) { #if defined(_MSC_VER) || defined(__MINGW32__) #include #else -#if !defined(__riscv) && !defined(__s390__) +#if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) || defined(__SSE3__) +#if !defined(__riscv) #include #endif #endif #endif #endif #endif +#endif #ifdef __riscv_v_intrinsic #include @@ -61,9 +51,13 @@ inline static int32_t vaddvq_s32(int32x4_t v) { #undef MIN #undef MAX + #define MIN(a, b) ((a) < (b) ? (a) : (b)) #define MAX(a, b) ((a) > (b) ? (a) : (b)) +#define UNUSED GGML_UNUSED + +// some compilers don't provide _mm256_set_m128i, e.g. gcc 7 #define MM256_SET_M128I(a, b) _mm256_insertf128_si256(_mm256_castsi128_si256(b), (a), 1) #if defined(__AVX__) || defined(__AVX2__) || defined(__AVX512F__) || defined(__SSSE3__) @@ -138,7 +132,7 @@ static inline __m256 sum_i16_pairs_float(const __m256i x) { } static inline __m256 mul_sum_us8_pairs_float(const __m256i ax, const __m256i sy) { -#if __AVXVNNI__ +#if defined(__AVXVNNI__) || defined(__AVX512VNNI__) const __m256i zero = _mm256_setzero_si256(); const __m256i summed_pairs = _mm256_dpbusd_epi32(zero, ax, sy); return _mm256_cvtepi32_ps(summed_pairs); @@ -284,8 +278,50 @@ static inline float hsum_float_4x4(const __m128 a, const __m128 b, const __m128 #if defined(__ARM_NEON) +#ifdef _MSC_VER + +#define ggml_vld1q_u32(w,x,y,z) { ((w) + ((uint64_t)(x) << 32)), ((y) + ((uint64_t)(z) << 32)) } + +#else + +#define ggml_vld1q_u32(w,x,y,z) { (w), (x), (y), (z) } + +#endif + #if !defined(__aarch64__) +// 64-bit compatibility + +// vaddvq_s16 +// vpaddq_s16 +// vpaddq_s32 +// vaddvq_s32 +// vaddvq_f32 +// vmaxvq_f32 +// vcvtnq_s32_f32 +// vzip1_u8 +// vzip2_u8 + +inline static int32_t vaddvq_s16(int16x8_t v) { + return + (int32_t)vgetq_lane_s16(v, 0) + (int32_t)vgetq_lane_s16(v, 1) + + (int32_t)vgetq_lane_s16(v, 2) + (int32_t)vgetq_lane_s16(v, 3) + + (int32_t)vgetq_lane_s16(v, 4) + (int32_t)vgetq_lane_s16(v, 5) + + (int32_t)vgetq_lane_s16(v, 6) + (int32_t)vgetq_lane_s16(v, 7); +} + +inline static int16x8_t vpaddq_s16(int16x8_t a, int16x8_t b) { + int16x4_t a0 = vpadd_s16(vget_low_s16(a), vget_high_s16(a)); + int16x4_t b0 = vpadd_s16(vget_low_s16(b), vget_high_s16(b)); + return vcombine_s16(a0, b0); +} + +inline static int32x4_t vpaddq_s32(int32x4_t a, int32x4_t b) { + int32x2_t a0 = vpadd_s32(vget_low_s32(a), vget_high_s32(a)); + int32x2_t b0 = vpadd_s32(vget_low_s32(b), vget_high_s32(b)); + return vcombine_s32(a0, b0); +} + inline static int32_t vaddvq_s32(int32x4_t v) { return vgetq_lane_s32(v, 0) + vgetq_lane_s32(v, 1) + vgetq_lane_s32(v, 2) + vgetq_lane_s32(v, 3); } @@ -311,7 +347,185 @@ inline static int32x4_t vcvtnq_s32_f32(float32x4_t v) { return res; } +inline static uint8x8_t vzip1_u8(uint8x8_t a, uint8x8_t b) { + uint8x8_t res; + + res[0] = a[0]; res[1] = b[0]; + res[2] = a[1]; res[3] = b[1]; + res[4] = a[2]; res[5] = b[2]; + res[6] = a[3]; res[7] = b[3]; + + return res; +} + +inline static uint8x8_t vzip2_u8(uint8x8_t a, uint8x8_t b) { + uint8x8_t res; + + res[0] = a[4]; res[1] = b[4]; + res[2] = a[5]; res[3] = b[5]; + res[4] = a[6]; res[5] = b[6]; + res[6] = a[7]; res[7] = b[7]; + + return res; +} + +// vld1q_s16_x2 +// vld1q_u8_x2 +// vld1q_u8_x4 +// vld1q_s8_x2 +// vld1q_s8_x4 +// TODO: double-check these work correctly + +typedef struct ggml_int16x8x2_t { + int16x8_t val[2]; +} ggml_int16x8x2_t; + +inline static ggml_int16x8x2_t ggml_vld1q_s16_x2(const int16_t * ptr) { + ggml_int16x8x2_t res; + + res.val[0] = vld1q_s16(ptr + 0); + res.val[1] = vld1q_s16(ptr + 8); + + return res; +} + +typedef struct ggml_uint8x16x2_t { + uint8x16_t val[2]; +} ggml_uint8x16x2_t; + +inline static ggml_uint8x16x2_t ggml_vld1q_u8_x2(const uint8_t * ptr) { + ggml_uint8x16x2_t res; + + res.val[0] = vld1q_u8(ptr + 0); + res.val[1] = vld1q_u8(ptr + 16); + + return res; +} + +typedef struct ggml_uint8x16x4_t { + uint8x16_t val[4]; +} ggml_uint8x16x4_t; + +inline static ggml_uint8x16x4_t ggml_vld1q_u8_x4(const uint8_t * ptr) { + ggml_uint8x16x4_t res; + + res.val[0] = vld1q_u8(ptr + 0); + res.val[1] = vld1q_u8(ptr + 16); + res.val[2] = vld1q_u8(ptr + 32); + res.val[3] = vld1q_u8(ptr + 48); + + return res; +} + +typedef struct ggml_int8x16x2_t { + int8x16_t val[2]; +} ggml_int8x16x2_t; + +inline static ggml_int8x16x2_t ggml_vld1q_s8_x2(const int8_t * ptr) { + ggml_int8x16x2_t res; + + res.val[0] = vld1q_s8(ptr + 0); + res.val[1] = vld1q_s8(ptr + 16); + + return res; +} + +typedef struct ggml_int8x16x4_t { + int8x16_t val[4]; +} ggml_int8x16x4_t; + +inline static ggml_int8x16x4_t ggml_vld1q_s8_x4(const int8_t * ptr) { + ggml_int8x16x4_t res; + + res.val[0] = vld1q_s8(ptr + 0); + res.val[1] = vld1q_s8(ptr + 16); + res.val[2] = vld1q_s8(ptr + 32); + res.val[3] = vld1q_s8(ptr + 48); + + return res; +} + +// NOTE: not tested +inline static int8x16_t ggml_vqtbl1q_s8(int8x16_t a, uint8x16_t b) { + int8x16_t res; + + res[ 0] = a[b[ 0]]; + res[ 1] = a[b[ 1]]; + res[ 2] = a[b[ 2]]; + res[ 3] = a[b[ 3]]; + res[ 4] = a[b[ 4]]; + res[ 5] = a[b[ 5]]; + res[ 6] = a[b[ 6]]; + res[ 7] = a[b[ 7]]; + res[ 8] = a[b[ 8]]; + res[ 9] = a[b[ 9]]; + res[10] = a[b[10]]; + res[11] = a[b[11]]; + res[12] = a[b[12]]; + res[13] = a[b[13]]; + res[14] = a[b[14]]; + res[15] = a[b[15]]; + + return res; +} + +// NOTE: not tested +inline static uint8x16_t ggml_vqtbl1q_u8(uint8x16_t a, uint8x16_t b) { + uint8x16_t res; + + res[ 0] = a[b[ 0]]; + res[ 1] = a[b[ 1]]; + res[ 2] = a[b[ 2]]; + res[ 3] = a[b[ 3]]; + res[ 4] = a[b[ 4]]; + res[ 5] = a[b[ 5]]; + res[ 6] = a[b[ 6]]; + res[ 7] = a[b[ 7]]; + res[ 8] = a[b[ 8]]; + res[ 9] = a[b[ 9]]; + res[10] = a[b[10]]; + res[11] = a[b[11]]; + res[12] = a[b[12]]; + res[13] = a[b[13]]; + res[14] = a[b[14]]; + res[15] = a[b[15]]; + + return res; +} + +#else + +#define ggml_int16x8x2_t int16x8x2_t +#define ggml_uint8x16x2_t uint8x16x2_t +#define ggml_uint8x16x4_t uint8x16x4_t +#define ggml_int8x16x2_t int8x16x2_t +#define ggml_int8x16x4_t int8x16x4_t + +#define ggml_vld1q_s16_x2 vld1q_s16_x2 +#define ggml_vld1q_u8_x2 vld1q_u8_x2 +#define ggml_vld1q_u8_x4 vld1q_u8_x4 +#define ggml_vld1q_s8_x2 vld1q_s8_x2 +#define ggml_vld1q_s8_x4 vld1q_s8_x4 +#define ggml_vqtbl1q_s8 vqtbl1q_s8 +#define ggml_vqtbl1q_u8 vqtbl1q_u8 + +#endif + +#if !defined(__ARM_FEATURE_DOTPROD) + +inline static int32x4_t ggml_vdotq_s32(int32x4_t acc, int8x16_t a, int8x16_t b) { + const int16x8_t p0 = vmull_s8(vget_low_s8 (a), vget_low_s8 (b)); + const int16x8_t p1 = vmull_s8(vget_high_s8(a), vget_high_s8(b)); + + return vaddq_s32(acc, vaddq_s32(vpaddlq_s16(p0), vpaddlq_s16(p1))); +} + +#else + +#define ggml_vdotq_s32(a, b, c) vdotq_s32(a, b, c) + #endif + #endif #if defined(__ARM_NEON) || defined(__wasm_simd128__) @@ -330,7 +544,7 @@ static const uint64_t table_b2b_1[1 << 8] = { B8(10, 00) }; // (!b) << 4 #endif // reference implementation for deterministic creation of model files -void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int k) { +void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int64_t k) { static const int qk = QK4_0; assert(k % qk == 0); @@ -367,11 +581,12 @@ void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict } } -void quantize_row_q4_0(const float * restrict x, void * restrict y, int k) { +void quantize_row_q4_0(const float * restrict x, void * restrict y, int64_t k) { quantize_row_q4_0_reference(x, y, k); } -void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict y, int k) { + +void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict y, int64_t k) { const int qk = QK4_1; assert(k % qk == 0); @@ -408,11 +623,11 @@ void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict } } -void quantize_row_q4_1(const float * restrict x, void * restrict y, int k) { +void quantize_row_q4_1(const float * restrict x, void * restrict y, int64_t k) { quantize_row_q4_1_reference(x, y, k); } -void quantize_row_q5_0_reference(const float * restrict x, block_q5_0 * restrict y, int k) { +void quantize_row_q5_0_reference(const float * restrict x, block_q5_0 * restrict y, int64_t k) { static const int qk = QK5_0; assert(k % qk == 0); @@ -456,11 +671,11 @@ void quantize_row_q5_0_reference(const float * restrict x, block_q5_0 * restrict } } -void quantize_row_q5_0(const float * restrict x, void * restrict y, int k) { +void quantize_row_q5_0(const float * restrict x, void * restrict y, int64_t k) { quantize_row_q5_0_reference(x, y, k); } -void quantize_row_q5_1_reference(const float * restrict x, block_q5_1 * restrict y, int k) { +void quantize_row_q5_1_reference(const float * restrict x, block_q5_1 * restrict y, int64_t k) { const int qk = QK5_1; assert(k % qk == 0); @@ -504,12 +719,12 @@ void quantize_row_q5_1_reference(const float * restrict x, block_q5_1 * restrict } } -void quantize_row_q5_1(const float * restrict x, void * restrict y, int k) { +void quantize_row_q5_1(const float * restrict x, void * restrict y, int64_t k) { quantize_row_q5_1_reference(x, y, k); } // reference implementation for deterministic creation of model files -void quantize_row_q8_0_reference(const float * restrict x, block_q8_0 * restrict y, int k) { +void quantize_row_q8_0_reference(const float * restrict x, block_q8_0 * restrict y, int64_t k) { assert(k % QK8_0 == 0); const int nb = k / QK8_0; @@ -534,7 +749,7 @@ void quantize_row_q8_0_reference(const float * restrict x, block_q8_0 * restrict } } -void quantize_row_q8_0(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q8_0(const float * restrict x, void * restrict vy, int64_t k) { assert(QK8_0 == 32); assert(k % QK8_0 == 0); const int nb = k / QK8_0; @@ -723,7 +938,7 @@ void quantize_row_q8_0(const float * restrict x, void * restrict vy, int k) { } // reference implementation for deterministic creation of model files -void quantize_row_q8_1_reference(const float * restrict x, block_q8_1 * restrict y, int k) { +void quantize_row_q8_1_reference(const float * restrict x, block_q8_1 * restrict y, int64_t k) { assert(QK8_1 == 32); assert(k % QK8_1 == 0); const int nb = k / QK8_1; @@ -739,7 +954,7 @@ void quantize_row_q8_1_reference(const float * restrict x, block_q8_1 * restrict const float d = amax / ((1 << 7) - 1); const float id = d ? 1.0f/d : 0.0f; - y[i].d = d; + y[i].d = GGML_FP32_TO_FP16(d); int sum = 0; @@ -754,11 +969,11 @@ void quantize_row_q8_1_reference(const float * restrict x, block_q8_1 * restrict sum += y[i].qs[QK8_1/2 + j]; } - y[i].s = sum*d; + y[i].s = GGML_FP32_TO_FP16(sum*d); } } -void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q8_1(const float * restrict x, void * restrict vy, int64_t k) { assert(k % QK8_1 == 0); const int nb = k / QK8_1; @@ -782,7 +997,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { const float d = amax / ((1 << 7) - 1); const float id = d ? 1.0f/d : 0.0f; - y[i].d = d; + y[i].d = GGML_FP32_TO_FP16(d); int32x4_t accv = vdupq_n_s32(0); @@ -798,7 +1013,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { accv = vaddq_s32(accv, vi); } - y[i].s = d * vaddvq_s32(accv); + y[i].s = GGML_FP32_TO_FP16(d * vaddvq_s32(accv)); } #elif defined(__wasm_simd128__) for (int i = 0; i < nb; i++) { @@ -821,7 +1036,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { const float d = amax / ((1 << 7) - 1); const float id = d ? 1.0f/d : 0.0f; - y[i].d = d; + y[i].d = GGML_FP32_TO_FP16(d); v128_t accv = wasm_i32x4_splat(0); @@ -837,10 +1052,11 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { accv = wasm_i32x4_add(accv, vi); } - y[i].s = d * (wasm_i32x4_extract_lane(accv, 0) + - wasm_i32x4_extract_lane(accv, 1) + - wasm_i32x4_extract_lane(accv, 2) + - wasm_i32x4_extract_lane(accv, 3)); + y[i].s = GGML_FP32_TO_FP16( + d * (wasm_i32x4_extract_lane(accv, 0) + + wasm_i32x4_extract_lane(accv, 1) + + wasm_i32x4_extract_lane(accv, 2) + + wasm_i32x4_extract_lane(accv, 3))); } #elif defined(__AVX2__) || defined(__AVX__) for (int i = 0; i < nb; i++) { @@ -865,7 +1081,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { // Quantize these floats const float d = maxScalar / 127.f; - y[i].d = d; + y[i].d = GGML_FP32_TO_FP16(d); const float id = ( maxScalar != 0.0f ) ? 127.f / maxScalar : 0.0f; const __m256 mul = _mm256_set1_ps( id ); @@ -889,7 +1105,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { #if defined(__AVX2__) // Compute the sum of the quants and set y[i].s - y[i].s = d * hsum_i32_8(_mm256_add_epi32(_mm256_add_epi32(i0, i1), _mm256_add_epi32(i2, i3))); + y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_8(_mm256_add_epi32(_mm256_add_epi32(i0, i1), _mm256_add_epi32(i2, i3)))); // Convert int32 to int16 i0 = _mm256_packs_epi32( i0, i1 ); // 0, 1, 2, 3, 8, 9, 10, 11, 4, 5, 6, 7, 12, 13, 14, 15 @@ -919,7 +1135,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { // Compute the sum of the quants and set y[i].s const __m128i s0 = _mm_add_epi32(_mm_add_epi32(ni0, ni1), _mm_add_epi32(ni2, ni3)); const __m128i s1 = _mm_add_epi32(_mm_add_epi32(ni4, ni5), _mm_add_epi32(ni6, ni7)); - y[i].s = d * hsum_i32_4(_mm_add_epi32(s0, s1)); + y[i].s = GGML_FP32_TO_FP16(d * hsum_i32_4(_mm_add_epi32(s0, s1))); // Convert int32 to int16 ni0 = _mm_packs_epi32( ni0, ni1 ); @@ -950,7 +1166,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { const float d = amax / ((1 << 7) - 1); const float id = d ? 1.0f/d : 0.0f; - y[i].d = d; + y[i].d = GGML_FP32_TO_FP16(d); vfloat32m4_t x0 = __riscv_vfmul_vf_f32m4(v_x, id, vl); @@ -967,7 +1183,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { // set y[i].s int sum = __riscv_vmv_x_s_i16m1_i16(vwrs); - y[i].s = sum*d; + y[i].s = GGML_FP32_TO_FP16(sum*d); } #else GGML_UNUSED(nb); @@ -976,7 +1192,7 @@ void quantize_row_q8_1(const float * restrict x, void * restrict vy, int k) { #endif } -void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int k) { +void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int64_t k) { static const int qk = QK4_0; assert(k % qk == 0); @@ -996,7 +1212,7 @@ void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int } } -void dequantize_row_q4_1(const block_q4_1 * restrict x, float * restrict y, int k) { +void dequantize_row_q4_1(const block_q4_1 * restrict x, float * restrict y, int64_t k) { static const int qk = QK4_1; assert(k % qk == 0); @@ -1017,7 +1233,7 @@ void dequantize_row_q4_1(const block_q4_1 * restrict x, float * restrict y, int } } -void dequantize_row_q5_0(const block_q5_0 * restrict x, float * restrict y, int k) { +void dequantize_row_q5_0(const block_q5_0 * restrict x, float * restrict y, int64_t k) { static const int qk = QK5_0; assert(k % qk == 0); @@ -1043,7 +1259,7 @@ void dequantize_row_q5_0(const block_q5_0 * restrict x, float * restrict y, int } } -void dequantize_row_q5_1(const block_q5_1 * restrict x, float * restrict y, int k) { +void dequantize_row_q5_1(const block_q5_1 * restrict x, float * restrict y, int64_t k) { static const int qk = QK5_1; assert(k % qk == 0); @@ -1070,7 +1286,7 @@ void dequantize_row_q5_1(const block_q5_1 * restrict x, float * restrict y, int } } -void dequantize_row_q8_0(const block_q8_0 * restrict x, float * restrict y, int k) { +void dequantize_row_q8_0(const block_q8_0 * restrict x, float * restrict y, int64_t k) { static const int qk = QK8_0; assert(k % qk == 0); @@ -1100,7 +1316,8 @@ static inline int nearest_int(float fval) { return (i & 0x007fffff) - 0x00400000; } -static float make_qx_quants(int n, int nmax, const float * restrict x, int8_t * restrict L, int rmse_type) { +static float make_qx_quants(int n, int nmax, const float * restrict x, int8_t * restrict L, int rmse_type, + const float * restrict qw) { float max = 0; float amax = 0; for (int i = 0; i < n; ++i) { @@ -1126,14 +1343,18 @@ static float make_qx_quants(int n, int nmax, const float * restrict x, int8_t * rmse_type = -rmse_type; return_early = true; } - int weight_type = rmse_type%2; float sumlx = 0; float suml2 = 0; +#ifdef HAVE_BUGGY_APPLE_LINKER + // use 'volatile' to prevent unroll and work around a bug in Apple ld64 1015.7 + for (volatile int i = 0; i < n; ++i) { +#else for (int i = 0; i < n; ++i) { +#endif int l = nearest_int(iscale * x[i]); l = MAX(-nmax, MIN(nmax-1, l)); L[i] = l + nmax; - float w = weight_type == 1 ? x[i] * x[i] : 1; + float w = qw ? qw[i] : rmse_type == 1 ? x[i] * x[i] : rmse_type == 2 ? 1 : rmse_type == 3 ? fabsf(x[i]) : sqrtf(fabsf(x[i])); sumlx += w*x[i]*l; suml2 += w*l*l; } @@ -1149,7 +1370,7 @@ static float make_qx_quants(int n, int nmax, const float * restrict x, int8_t * for (int i = 0; i < n; ++i) { int l = nearest_int(iscale * x[i]); l = MAX(-nmax, MIN(nmax-1, l)); - float w = weight_type == 1 ? x[i] * x[i] : 1; + float w = qw ? qw[i] : rmse_type == 1 ? x[i] * x[i] : rmse_type == 2 ? 1 : rmse_type == 3 ? fabsf(x[i]) : sqrtf(fabsf(x[i])); sumlx += w*x[i]*l; suml2 += w*l*l; } @@ -1273,7 +1494,12 @@ static float make_qkx2_quants(int n, int nmax, const float * restrict x, const f float max = x[0]; float sum_w = weights[0]; float sum_x = sum_w * x[0]; +#ifdef HAVE_BUGGY_APPLE_LINKER + // use 'volatile' to prevent unroll and work around a bug in Apple ld64 1015.7 + for (volatile int i = 1; i < n; ++i) { +#else for (int i = 1; i < n; ++i) { +#endif if (x[i] < min) min = x[i]; if (x[i] > max) max = x[i]; float w = weights[i]; @@ -1355,7 +1581,7 @@ static inline void get_scale_min_k4(int j, const uint8_t * restrict q, uint8_t * //========================- 2-bit (de)-quantization -void quantize_row_q2_K_reference(const float * restrict x, block_q2_K * restrict y, int k) { +void quantize_row_q2_K_reference(const float * restrict x, block_q2_K * restrict y, int64_t k) { assert(k % QK_K == 0); const int nb = k / QK_K; @@ -1432,7 +1658,7 @@ void quantize_row_q2_K_reference(const float * restrict x, block_q2_K * restrict } } -void dequantize_row_q2_K(const block_q2_K * restrict x, float * restrict y, int k) { +void dequantize_row_q2_K(const block_q2_K * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); const int nb = k / QK_K; @@ -1478,64 +1704,322 @@ void dequantize_row_q2_K(const block_q2_K * restrict x, float * restrict y, int } } -void quantize_row_q2_K(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q2_K(const float * restrict x, void * restrict vy, int64_t k) { quantize_row_q2_K_reference(x, vy, k); } -size_t ggml_quantize_q2_K(const float * restrict src, void * restrict dst, int n, int k, int64_t * restrict hist) { - (void)hist; // TODO: collect histograms - - for (int j = 0; j < n; j += k) { - block_q2_K * restrict y = (block_q2_K *)dst + j/QK_K; - quantize_row_q2_K_reference(src + j, y, k); +static float make_qkx3_quants(int n, int nmax, const float * restrict x, const float * restrict weights, + uint8_t * restrict L, float * restrict the_min, uint8_t * restrict Laux, + float rmin, float rdelta, int nstep, bool use_mad) { + float min = x[0]; + float max = x[0]; + float sum_w = weights ? weights[0] : x[0]*x[0]; + float sum_x = sum_w * x[0]; +#ifdef HAVE_BUGGY_APPLE_LINKER + // use 'volatile' to prevent unroll and work around a bug in Apple ld64 1015.7 + for (volatile int i = 1; i < n; ++i) { +#else + for (int i = 1; i < n; ++i) { +#endif + if (x[i] < min) min = x[i]; + if (x[i] > max) max = x[i]; + float w = weights ? weights[i] : x[i]*x[i]; + sum_w += w; + sum_x += w * x[i]; + } + if (min > 0) { + min = 0; + } + if (max <= min) { + memset(L, 0, n); + *the_min = -min; + return 0.f; + } + float iscale = nmax/(max - min); + float scale = 1/iscale; + float best_mad = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale*(x[i] - min)); + L[i] = MAX(0, MIN(nmax, l)); + float diff = scale * L[i] + min - x[i]; + diff = use_mad ? fabsf(diff) : diff*diff; + float w = weights ? weights[i] : x[i]*x[i]; + best_mad += w * diff; + } + if (nstep < 1) { + *the_min = -min; + return scale; + } + for (int is = 0; is <= nstep; ++is) { + iscale = (rmin + rdelta*is + nmax)/(max - min); + float sum_l = 0, sum_l2 = 0, sum_xl = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale*(x[i] - min)); + l = MAX(0, MIN(nmax, l)); + Laux[i] = l; + float w = weights ? weights[i] : x[i]*x[i]; + sum_l += w*l; + sum_l2 += w*l*l; + sum_xl += w*l*x[i]; + } + float D = sum_w * sum_l2 - sum_l * sum_l; + if (D > 0) { + float this_scale = (sum_w * sum_xl - sum_x * sum_l)/D; + float this_min = (sum_l2 * sum_x - sum_l * sum_xl)/D; + if (this_min > 0) { + this_min = 0; + this_scale = sum_xl / sum_l2; + } + float mad = 0; + for (int i = 0; i < n; ++i) { + float diff = this_scale * Laux[i] + this_min - x[i]; + diff = use_mad ? fabsf(diff) : diff*diff; + float w = weights ? weights[i] : x[i]*x[i]; + mad += w * diff; + } + if (mad < best_mad) { + for (int i = 0; i < n; ++i) { + L[i] = Laux[i]; + } + best_mad = mad; + scale = this_scale; + min = this_min; + } + } } - return (n/QK_K*sizeof(block_q2_K)); + *the_min = -min; + return scale; } -//========================= 3-bit (de)-quantization +static float make_qp_quants(int n, int nmax, const float * restrict x, uint8_t * restrict L, const float * quant_weights) { + float max = 0; + for (int i = 0; i < n; ++i) { + max = MAX(max, x[i]); + } + if (!max) { // all zero + for (int i = 0; i < n; ++i) { L[i] = 0; } + return 0.f; + } + float iscale = nmax / max; + for (int i = 0; i < n; ++i) { + L[i] = nearest_int(iscale * x[i]); + } + float scale = 1/iscale; + float best_mse = 0; + for (int i = 0; i < n; ++i) { + float diff = x[i] - scale*L[i]; + float w = quant_weights[i]; + best_mse += w*diff*diff; + } + for (int is = -4; is <= 4; ++is) { + if (is == 0) continue; + float iscale_is = (0.1f*is + nmax)/max; + float scale_is = 1/iscale_is; + float mse = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale_is*x[i]); + l = MIN(nmax, l); + float diff = x[i] - scale_is*l; + float w = quant_weights[i]; + mse += w*diff*diff; + } + if (mse < best_mse) { + best_mse = mse; + iscale = iscale_is; + } + } + float sumlx = 0; + float suml2 = 0; + for (int i = 0; i < n; ++i) { + int l = nearest_int(iscale * x[i]); + l = MIN(nmax, l); + L[i] = l; + float w = quant_weights[i]; + sumlx += w*x[i]*l; + suml2 += w*l*l; + } + for (int itry = 0; itry < 5; ++itry) { + int n_changed = 0; + for (int i = 0; i < n; ++i) { + float w = quant_weights[i]; + float slx = sumlx - w*x[i]*L[i]; + float sl2 = suml2 - w*L[i]*L[i]; + if (slx > 0 && sl2 > 0) { + int new_l = nearest_int(x[i] * sl2 / slx); + new_l = MIN(nmax, new_l); + if (new_l != L[i]) { + slx += w*x[i]*new_l; + sl2 += w*new_l*new_l; + if (slx*slx*suml2 > sumlx*sumlx*sl2) { + L[i] = new_l; sumlx = slx; suml2 = sl2; + ++n_changed; + } + } + } + } + if (!n_changed) { + break; + } + } + return sumlx / suml2; +} -void quantize_row_q3_K_reference(const float * restrict x, block_q3_K * restrict y, int k) { +static void quantize_row_q2_K_impl(const float * restrict x, block_q2_K * restrict y, int k, const float * restrict quant_weights) { + GGML_ASSERT(quant_weights); assert(k % QK_K == 0); const int nb = k / QK_K; + const bool requantize = true; - int8_t L[QK_K]; - float scales[QK_K / 16]; + uint8_t L[QK_K]; + uint8_t Laux[16]; + float mins[QK_K/16]; + float scales[QK_K/16]; + float sw[QK_K/16]; + float weight[16]; + uint8_t Ls[QK_K/16], Lm[QK_K/16]; for (int i = 0; i < nb; i++) { - - float max_scale = 0; - float amax = 0; + memset(sw, 0, QK_K/16*sizeof(float)); + float sumx2 = 0; + for (int j = 0; j < QK_K; ++j) sumx2 += x[j]*x[j]; + float sigma2 = sumx2/QK_K; for (int j = 0; j < QK_K/16; ++j) { - scales[j] = make_q3_quants(16, 4, x + 16*j, L + 16*j, true); - float scale = fabsf(scales[j]); - if (scale > amax) { - amax = scale; max_scale = scales[j]; - } + const float * restrict qw = quant_weights + QK_K * i + 16*j; + for (int l = 0; l < 16; ++l) weight[l] = qw[l] * sqrtf(sigma2 + x[16*j + l]*x[16*j + l]); + for (int l = 0; l < QK_K/16; ++l) sw[j] += weight[l]; + scales[j] = make_qkx3_quants(16, 3, x + 16*j, weight, L + 16*j, &mins[j], Laux, -0.9f, 0.05f, 36, false); } -#if QK_K == 256 - memset(y[i].scales, 0, 12); + float dm, mm; +#if QK_K == 64 + float max_scale = 0, max_min = 0; + for (int j = 0; j < QK_K/16; ++j) { + max_scale = MAX(max_scale, scales[j]); + max_min = MAX(max_min, mins[j]); + } + dm = max_scale/15; + mm = max_min/15; if (max_scale) { - float iscale = -32.f/max_scale; + float id = 1/dm; for (int j = 0; j < QK_K/16; ++j) { - int8_t l = nearest_int(iscale*scales[j]); - l = MAX(-32, MIN(31, l)) + 32; - if (j < 8) { - y[i].scales[j] = l & 0xF; - } else { - y[i].scales[j-8] |= ((l & 0xF) << 4); - } - l >>= 4; - y[i].scales[j%4 + 8] |= (l << (2*(j/4))); + int l = nearest_int(id*scales[j]); + Ls[j] = MAX(0, MIN(15, l)); } - y[i].d = GGML_FP32_TO_FP16(1/iscale); } else { - y[i].d = GGML_FP32_TO_FP16(0.f); + memset(Ls, 0, QK_K/16); } + if (max_min) { + float id = 1/mm; + for (int j = 0; j < QK_K/16; ++j) { + int l = nearest_int(id*mins[j]); + Lm[j] = MAX(0, MIN(15, l)); + } + } else { + memset(Lm, 0, QK_K/16); + } +#else + dm = make_qp_quants(QK_K/16, 15, scales, Ls, sw); + mm = make_qp_quants(QK_K/16, 15, mins, Lm, sw); +#endif + y[i].d = GGML_FP32_TO_FP16(dm); + y[i].dmin = GGML_FP32_TO_FP16(mm); + dm = GGML_FP16_TO_FP32(y[i].d); + mm = GGML_FP16_TO_FP32(y[i].dmin); - int8_t sc; for (int j = 0; j < QK_K/16; ++j) { - sc = j < 8 ? y[i].scales[j] & 0xF : y[i].scales[j-8] >> 4; + y[i].scales[j] = Ls[j] | (Lm[j] << 4); + } + + if (requantize) { + for (int j = 0; j < QK_K/16; ++j) { + const float d = dm * (y[i].scales[j] & 0xF); + if (!d) continue; + const float m = mm * (y[i].scales[j] >> 4); + for (int ii = 0; ii < 16; ++ii) { + int l = nearest_int((x[16*j + ii] + m)/d); + l = MAX(0, MIN(3, l)); + L[16*j + ii] = l; + } + } + } + +#if QK_K == 256 + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) { + y[i].qs[j/4 + l] = L[j + l] | (L[j + l + 32] << 2) | (L[j + l + 64] << 4) | (L[j + l + 96] << 6); + } + } +#else + for (int l = 0; l < 16; ++l) { + y[i].qs[l] = L[l] | (L[l + 16] << 2) | (L[l + 32] << 4) | (L[l + 48] << 6); + } +#endif + + x += QK_K; + + } +} + +size_t quantize_q2_K(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + size_t row_size = ggml_row_size(GGML_TYPE_Q2_K, n_per_row); + if (!quant_weights) { + quantize_row_q2_K_reference(src, dst, (int64_t)nrow*n_per_row); + } + else { + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q2_K_impl(src, (block_q2_K*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + } + return nrow * row_size; +} + +//========================= 3-bit (de)-quantization + +void quantize_row_q3_K_reference(const float * restrict x, block_q3_K * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int nb = k / QK_K; + + int8_t L[QK_K]; + float scales[QK_K / 16]; + + for (int i = 0; i < nb; i++) { + + float max_scale = 0; + float amax = 0; + for (int j = 0; j < QK_K/16; ++j) { + scales[j] = make_q3_quants(16, 4, x + 16*j, L + 16*j, true); + float scale = fabsf(scales[j]); + if (scale > amax) { + amax = scale; max_scale = scales[j]; + } + } + +#if QK_K == 256 + memset(y[i].scales, 0, 12); + if (max_scale) { + float iscale = -32.f/max_scale; + for (int j = 0; j < QK_K/16; ++j) { + int8_t l = nearest_int(iscale*scales[j]); + l = MAX(-32, MIN(31, l)) + 32; + if (j < 8) { + y[i].scales[j] = l & 0xF; + } else { + y[i].scales[j-8] |= ((l & 0xF) << 4); + } + l >>= 4; + y[i].scales[j%4 + 8] |= (l << (2*(j/4))); + } + y[i].d = GGML_FP32_TO_FP16(1/iscale); + } else { + y[i].d = GGML_FP32_TO_FP16(0.f); + } + + int8_t sc; + for (int j = 0; j < QK_K/16; ++j) { + sc = j < 8 ? y[i].scales[j] & 0xF : y[i].scales[j-8] >> 4; sc = (sc | (((y[i].scales[8 + j%4] >> (2*(j/4))) & 3) << 4)) - 32; float d = GGML_FP16_TO_FP32(y[i].d) * sc; if (!d) { @@ -1608,7 +2092,7 @@ void quantize_row_q3_K_reference(const float * restrict x, block_q3_K * restrict } #if QK_K == 256 -void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int k) { +void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); const int nb = k / QK_K; @@ -1658,7 +2142,7 @@ void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int } } #else -void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int k) { +void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); assert(QK_K == 64); const int nb = k / QK_K; @@ -1691,23 +2175,118 @@ void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int } #endif -void quantize_row_q3_K(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q3_K(const float * restrict x, void * restrict vy, int64_t k) { quantize_row_q3_K_reference(x, vy, k); } -size_t ggml_quantize_q3_K(const float * restrict src, void * restrict dst, int n, int k, int64_t * restrict hist) { - (void)hist; // TODO: collect histograms +static void quantize_row_q3_K_impl(const float * restrict x, block_q3_K * restrict y, int64_t n_per_row, const float * restrict quant_weights) { +#if QK_K != 256 + (void)quant_weights; + quantize_row_q3_K_reference(x, y, n_per_row); +#else + assert(n_per_row % QK_K == 0); + const int nb = n_per_row / QK_K; + + int8_t L[QK_K]; + float scales[QK_K / 16]; + float weight[16]; + float sw[QK_K / 16]; + int8_t Ls[QK_K / 16]; + + for (int i = 0; i < nb; i++) { + + float sumx2 = 0; + for (int j = 0; j < QK_K; ++j) sumx2 += x[j]*x[j]; + float sigma2 = 2*sumx2/QK_K; + + for (int j = 0; j < QK_K/16; ++j) { + if (quant_weights) { + const float * qw = quant_weights ? quant_weights + QK_K * i + 16*j : NULL; + for (int l = 0; l < 16; ++l) weight[l] = qw[l] * sqrtf(sigma2 + x[16*j+l]*x[16*j+l]); + } else { + for (int l = 0; l < 16; ++l) weight[l] = x[16*j+l]*x[16*j+l]; + } + float sumw = 0; + for (int l = 0; l < 16; ++l) sumw += weight[l]; + sw[j] = sumw; + + scales[j] = make_qx_quants(16, 4, x + 16*j, L + 16*j, 1, weight); + + } + + memset(y[i].scales, 0, 12); + + float d_block = make_qx_quants(QK_K/16, 32, scales, Ls, 1, sw); + for (int j = 0; j < QK_K/16; ++j) { + int l = Ls[j]; + if (j < 8) { + y[i].scales[j] = l & 0xF; + } else { + y[i].scales[j-8] |= ((l & 0xF) << 4); + } + l >>= 4; + y[i].scales[j%4 + 8] |= (l << (2*(j/4))); + } + y[i].d = GGML_FP32_TO_FP16(d_block); + + int8_t sc; + for (int j = 0; j < QK_K/16; ++j) { + sc = j < 8 ? y[i].scales[j] & 0xF : y[i].scales[j-8] >> 4; + sc = (sc | (((y[i].scales[8 + j%4] >> (2*(j/4))) & 3) << 4)) - 32; + float d = GGML_FP16_TO_FP32(y[i].d) * sc; + if (!d) { + continue; + } + for (int ii = 0; ii < 16; ++ii) { + int l = nearest_int(x[16*j + ii]/d); + l = MAX(-4, MIN(3, l)); + L[16*j + ii] = l + 4; + } + } + + memset(y[i].hmask, 0, QK_K/8); + // We put the high-bit for the 1st 8 quants into bit 0, the next 8 into bit 1, etc. + int m = 0; + uint8_t hm = 1; + for (int j = 0; j < QK_K; ++j) { + if (L[j] > 3) { + y[i].hmask[m] |= hm; + L[j] -= 4; + } + if (++m == QK_K/8) { + m = 0; hm <<= 1; + } + } + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) { + y[i].qs[j/4 + l] = L[j + l] | (L[j + l + 32] << 2) | (L[j + l + 64] << 4) | (L[j + l + 96] << 6); + } + } + + x += QK_K; + } +#endif +} - for (int j = 0; j < n; j += k) { - block_q3_K * restrict y = (block_q3_K *)dst + j/QK_K; - quantize_row_q3_K_reference(src + j, y, k); +size_t quantize_q3_K(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + size_t row_size = ggml_row_size(GGML_TYPE_Q3_K, n_per_row); + if (!quant_weights) { + quantize_row_q3_K_reference(src, dst, (int64_t)nrow*n_per_row); } - return (n/QK_K*sizeof(block_q3_K)); + else { + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q3_K_impl(src, (block_q3_K*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + } + return nrow * row_size; } // ====================== 4-bit (de)-quantization -void quantize_row_q4_K_reference(const float * restrict x, block_q4_K * restrict y, int k) { +void quantize_row_q4_K_reference(const float * restrict x, block_q4_K * restrict y, int64_t k) { assert(k % QK_K == 0); const int nb = k / QK_K; @@ -1814,7 +2393,7 @@ void quantize_row_q4_K_reference(const float * restrict x, block_q4_K * restrict } } -void dequantize_row_q4_K(const block_q4_K * restrict x, float * restrict y, int k) { +void dequantize_row_q4_K(const block_q4_K * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); const int nb = k / QK_K; @@ -1853,28 +2432,111 @@ void dequantize_row_q4_K(const block_q4_K * restrict x, float * restrict y, int } } -void quantize_row_q4_K(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q4_K(const float * restrict x, void * restrict vy, int64_t k) { assert(k % QK_K == 0); block_q4_K * restrict y = vy; quantize_row_q4_K_reference(x, y, k); } -size_t ggml_quantize_q4_K(const float * restrict src, void * restrict dst, int n, int k, int64_t * restrict hist) { - assert(k % QK_K == 0); - (void)hist; // TODO: collect histograms +static void quantize_row_q4_K_impl(const float * restrict x, block_q4_K * restrict y, int64_t n_per_row, const float * quant_weights) { +#if QK_K != 256 + (void)quant_weights; + quantize_row_q4_K_reference(x, y, n_per_row); +#else + assert(n_per_row % QK_K == 0); + const int64_t nb = n_per_row / QK_K; + + uint8_t L[QK_K]; + uint8_t Laux[32]; + uint8_t Ls[QK_K/32]; + uint8_t Lm[QK_K/32]; + float weights[32]; + float sw[QK_K/32]; + float mins[QK_K/32]; + float scales[QK_K/32]; + + for (int i = 0; i < nb; i++) { + + float sum_x2 = 0; + for (int l = 0; l < QK_K; ++l) sum_x2 += x[l] * x[l]; + float sigma2 = 2*sum_x2/QK_K; + float av_x = sqrtf(sigma2); + + for (int j = 0; j < QK_K/32; ++j) { + if (quant_weights) { + const float * qw = quant_weights + QK_K*i + 32*j; + for (int l = 0; l < 32; ++l) weights[l] = qw[l] * sqrtf(sigma2 + x[32*j + l]*x[32*j + l]); + } else { + for (int l = 0; l < 32; ++l) weights[l] = av_x + fabsf(x[32*j + l]); + } + float sumw = 0; + for (int l = 0; l < 32; ++l) sumw += weights[l]; + sw[j] = sumw; + scales[j] = make_qkx3_quants(32, 15, x + 32*j, weights, L + 32*j, &mins[j], Laux, -0.9f, 0.05f, 36, false); + } + + float d_block = make_qp_quants(QK_K/32, 63, scales, Ls, sw); + float m_block = make_qp_quants(QK_K/32, 63, mins, Lm, sw); + for (int j = 0; j < QK_K/32; ++j) { + uint8_t ls = Ls[j]; + uint8_t lm = Lm[j]; + if (j < 4) { + y[i].scales[j] = ls; + y[i].scales[j+4] = lm; + } else { + y[i].scales[j+4] = (ls & 0xF) | ((lm & 0xF) << 4); + y[i].scales[j-4] |= ((ls >> 4) << 6); + y[i].scales[j-0] |= ((lm >> 4) << 6); + } + } + y[i].d = GGML_FP32_TO_FP16(d_block); + y[i].dmin = GGML_FP32_TO_FP16(m_block); + + uint8_t sc, m; + for (int j = 0; j < QK_K/32; ++j) { + get_scale_min_k4(j, y[i].scales, &sc, &m); + const float d = GGML_FP16_TO_FP32(y[i].d) * sc; + if (!d) continue; + const float dm = GGML_FP16_TO_FP32(y[i].dmin) * m; + for (int ii = 0; ii < 32; ++ii) { + int l = nearest_int((x[32*j + ii] + dm)/d); + l = MAX(0, MIN(15, l)); + L[32*j + ii] = l; + } + } + uint8_t * q = y[i].qs; + for (int j = 0; j < QK_K; j += 64) { + for (int l = 0; l < 32; ++l) q[l] = L[j + l] | (L[j + l + 32] << 4); + q += 32; + } + + x += QK_K; + + } +#endif +} - for (int j = 0; j < n; j += k) { - block_q4_K * restrict y = (block_q4_K *)dst + j/QK_K; - quantize_row_q4_K_reference(src + j, y, k); +size_t quantize_q4_K(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + size_t row_size = ggml_row_size(GGML_TYPE_Q4_K, n_per_row); + if (!quant_weights) { + quantize_row_q4_K_reference(src, dst, (int64_t)nrow*n_per_row); + } + else { + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q4_K_impl(src, (block_q4_K*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } } - return (n/QK_K*sizeof(block_q4_K)); + return nrow * row_size; } // ====================== 5-bit (de)-quantization -void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict y, int k) { +void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict y, int64_t k) { assert(k % QK_K == 0); - const int nb = k / QK_K; + const int64_t nb = k / QK_K; #if QK_K == 256 uint8_t L[QK_K]; @@ -1965,7 +2627,7 @@ void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict #else float max_scale = 0, amax = 0; for (int j = 0; j < QK_K/16; ++j) { - scales[j] = make_qx_quants(16, 16, x + 16*j, L + 16*j, 1); + scales[j] = make_qx_quants(16, 16, x + 16*j, L + 16*j, 1, NULL); float abs_scale = fabsf(scales[j]); if (abs_scale > amax) { amax = abs_scale; @@ -2014,9 +2676,9 @@ void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict } } -void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int k) { +void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); - const int nb = k / QK_K; + const int64_t nb = k / QK_K; for (int i = 0; i < nb; i++) { @@ -2059,78 +2721,181 @@ void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int } } -void quantize_row_q5_K(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q5_K(const float * restrict x, void * restrict vy, int64_t k) { assert(k % QK_K == 0); block_q5_K * restrict y = vy; quantize_row_q5_K_reference(x, y, k); } -size_t ggml_quantize_q5_K(const float * restrict src, void * restrict dst, int n, int k, int64_t * restrict hist) { - assert(k % QK_K == 0); - (void)hist; // TODO: collect histograms - - for (int j = 0; j < n; j += k) { - block_q5_K * restrict y = (block_q5_K *)dst + j/QK_K; - quantize_row_q5_K_reference(src + j, y, k); - } - return (n/QK_K*sizeof(block_q5_K)); -} - -// ====================== 6-bit (de)-quantization - -void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int k) { - assert(k % QK_K == 0); - const int nb = k / QK_K; +static void quantize_row_q5_K_impl(const float * restrict x, block_q5_K * restrict y, int64_t n_per_row, const float * quant_weights) { +#if QK_K != 256 + (void)quant_weights; + quantize_row_q5_K_reference(x, y, n_per_row); +#else + assert(n_per_row % QK_K == 0); + const int64_t nb = n_per_row / QK_K; - int8_t L[QK_K]; - float scales[QK_K/16]; + uint8_t L[QK_K]; + uint8_t Laux[32]; + uint8_t Ls[QK_K/32]; + uint8_t Lm[QK_K/32]; + float mins[QK_K/32]; + float scales[QK_K/32]; + float sw[QK_K/32]; + float weights[32]; for (int i = 0; i < nb; i++) { - float max_scale = 0; - float max_abs_scale = 0; - - for (int ib = 0; ib < QK_K/16; ++ib) { - - const float scale = make_qx_quants(16, 32, x + 16*ib, L + 16*ib, 1); - scales[ib] = scale; + float sum_x2 = 0; + for (int l = 0; l < QK_K; ++l) sum_x2 += x[l] * x[l]; + float sigma2 = 2*sum_x2/QK_K; + float av_x = sqrtf(sigma2); - const float abs_scale = fabsf(scale); - if (abs_scale > max_abs_scale) { - max_abs_scale = abs_scale; - max_scale = scale; + for (int j = 0; j < QK_K/32; ++j) { + if (quant_weights) { + const float * qw = quant_weights + QK_K*i + 32*j; + for (int l = 0; l < 32; ++l) weights[l] = qw[l] * sqrtf(sigma2 + x[32*j + l]*x[32*j + l]); + } else { + for (int l = 0; l < 32; ++l) weights[l] = av_x + fabsf(x[32*j + l]); } + float sumw = 0; + for (int l = 0; l < 32; ++l) sumw += weights[l]; + sw[j] = sumw; + scales[j] = make_qkx3_quants(32, 31, x + 32*j, weights, L + 32*j, &mins[j], Laux, -0.9f, 0.05f, 36, false); } - if (!max_abs_scale) { - memset(&y[i], 0, sizeof(block_q6_K)); - y[i].d = GGML_FP32_TO_FP16(0.f); - x += QK_K; - continue; - } + float d_block = make_qp_quants(QK_K/32, 63, scales, Ls, sw); + float m_block = make_qp_quants(QK_K/32, 63, mins, Lm, sw); - float iscale = -128.f/max_scale; - y[i].d = GGML_FP32_TO_FP16(1/iscale); - for (int ib = 0; ib < QK_K/16; ++ib) { - y[i].scales[ib] = MIN(127, nearest_int(iscale*scales[ib])); + for (int j = 0; j < QK_K/32; ++j) { + uint8_t ls = Ls[j]; + uint8_t lm = Lm[j]; + ls = MIN(63, ls); + lm = MIN(63, lm); + if (j < 4) { + y[i].scales[j] = ls; + y[i].scales[j+4] = lm; + } else { + y[i].scales[j+4] = (ls & 0xF) | ((lm & 0xF) << 4); + y[i].scales[j-4] |= ((ls >> 4) << 6); + y[i].scales[j-0] |= ((lm >> 4) << 6); + } } + y[i].d = GGML_FP32_TO_FP16(d_block); + y[i].dmin = GGML_FP32_TO_FP16(m_block); - for (int j = 0; j < QK_K/16; ++j) { - float d = GGML_FP16_TO_FP32(y[i].d) * y[i].scales[j]; - if (!d) { - continue; - } - for (int ii = 0; ii < 16; ++ii) { - int l = nearest_int(x[16*j + ii]/d); - l = MAX(-32, MIN(31, l)); - L[16*j + ii] = l + 32; + uint8_t sc, m; + for (int j = 0; j < QK_K/32; ++j) { + get_scale_min_k4(j, y[i].scales, &sc, &m); + const float d = GGML_FP16_TO_FP32(y[i].d) * sc; + if (!d) continue; + const float dm = GGML_FP16_TO_FP32(y[i].dmin) * m; + for (int ii = 0; ii < 32; ++ii) { + int l = nearest_int((x[32*j + ii] + dm)/d); + l = MAX(0, MIN(31, l)); + L[32*j + ii] = l; } } - uint8_t * restrict ql = y[i].ql; uint8_t * restrict qh = y[i].qh; -#if QK_K == 256 + uint8_t * restrict ql = y[i].qs; + memset(qh, 0, QK_K/8); + + uint8_t m1 = 1, m2 = 2; + for (int n = 0; n < QK_K; n += 64) { + for (int j = 0; j < 32; ++j) { + int l1 = L[n + j]; + if (l1 > 15) { + l1 -= 16; qh[j] |= m1; + } + int l2 = L[n + j + 32]; + if (l2 > 15) { + l2 -= 16; qh[j] |= m2; + } + ql[j] = l1 | (l2 << 4); + } + m1 <<= 2; m2 <<= 2; + ql += 32; + } + + x += QK_K; + + } +#endif +} + +size_t quantize_q5_K(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + size_t row_size = ggml_row_size(GGML_TYPE_Q5_K, n_per_row); + if (!quant_weights) { + quantize_row_q5_K_reference(src, dst, (int64_t)nrow*n_per_row); + } + else { + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q5_K_impl(src, (block_q5_K*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + } + return nrow * row_size; +} + +// ====================== 6-bit (de)-quantization + +void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; + + int8_t L[QK_K]; + float scales[QK_K/16]; + + for (int i = 0; i < nb; i++) { + + float max_scale = 0; + float max_abs_scale = 0; + + for (int ib = 0; ib < QK_K/16; ++ib) { + + const float scale = make_qx_quants(16, 32, x + 16*ib, L + 16*ib, 1, NULL); + scales[ib] = scale; + + const float abs_scale = fabsf(scale); + if (abs_scale > max_abs_scale) { + max_abs_scale = abs_scale; + max_scale = scale; + } + + } + + if (!max_abs_scale) { + memset(&y[i], 0, sizeof(block_q6_K)); + y[i].d = GGML_FP32_TO_FP16(0.f); + x += QK_K; + continue; + } + + float iscale = -128.f/max_scale; + y[i].d = GGML_FP32_TO_FP16(1/iscale); + for (int ib = 0; ib < QK_K/16; ++ib) { + y[i].scales[ib] = MIN(127, nearest_int(iscale*scales[ib])); + } + + for (int j = 0; j < QK_K/16; ++j) { + float d = GGML_FP16_TO_FP32(y[i].d) * y[i].scales[j]; + if (!d) { + continue; + } + for (int ii = 0; ii < 16; ++ii) { + int l = nearest_int(x[16*j + ii]/d); + l = MAX(-32, MIN(31, l)); + L[16*j + ii] = l + 32; + } + } + + uint8_t * restrict ql = y[i].ql; + uint8_t * restrict qh = y[i].qh; +#if QK_K == 256 for (int j = 0; j < QK_K; j += 128) { for (int l = 0; l < 32; ++l) { const uint8_t q1 = L[j + l + 0] & 0xF; @@ -2160,9 +2925,9 @@ void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict } } -void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int k) { +void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int64_t k) { assert(k % QK_K == 0); - const int nb = k / QK_K; + const int64_t nb = k / QK_K; for (int i = 0; i < nb; i++) { @@ -2207,459 +2972,815 @@ void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int } } -void quantize_row_q6_K(const float * restrict x, void * restrict vy, int k) { +void quantize_row_q6_K(const float * restrict x, void * restrict vy, int64_t k) { assert(k % QK_K == 0); block_q6_K * restrict y = vy; quantize_row_q6_K_reference(x, y, k); } -size_t ggml_quantize_q6_K(const float * src, void * dst, int n, int k, int64_t * hist) { - assert(k % QK_K == 0); - (void)hist; // TODO: collect histograms +static void quantize_row_q6_K_impl(const float * restrict x, block_q6_K * restrict y, int64_t n_per_row, const float * quant_weights) { +#if QK_K != 256 + (void)quant_weights; + quantize_row_q6_K_reference(x, y, n_per_row); +#else + assert(n_per_row % QK_K == 0); + const int64_t nb = n_per_row / QK_K; - for (int j = 0; j < n; j += k) { - block_q6_K * restrict y = (block_q6_K *)dst + j/QK_K; - quantize_row_q6_K_reference(src + j, y, k); - } - return (n/QK_K*sizeof(block_q6_K)); -} + int8_t L[QK_K]; + float scales[QK_K/16]; + //float weights[16]; -//===================================== Q8_K ============================================== + for (int i = 0; i < nb; i++) { -void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int k) { - assert(k % QK_K == 0); - const int nb = k / QK_K; + //float sum_x2 = 0; + //for (int j = 0; j < QK_K; ++j) sum_x2 += x[j]*x[j]; + //float sigma2 = sum_x2/QK_K; - for (int i = 0; i < nb; i++) { + float max_scale = 0; + float max_abs_scale = 0; - float max = 0; - float amax = 0; - for (int j = 0; j < QK_K; ++j) { - float ax = fabsf(x[j]); - if (ax > amax) { - amax = ax; max = x[j]; + for (int ib = 0; ib < QK_K/16; ++ib) { + + float scale; + if (quant_weights) { + const float * qw = quant_weights + QK_K*i + 16*ib; + //for (int j = 0; j < 16; ++j) weights[j] = qw[j] * sqrtf(sigma2 + x[16*ib + j]*x[16*ib + j]); + //scale = make_qx_quants(16, 32, x + 16*ib, L + 16*ib, 1, weights); + scale = make_qx_quants(16, 32, x + 16*ib, L + 16*ib, 1, qw); + } else { + scale = make_qx_quants(16, 32, x + 16*ib, L + 16*ib, 1, NULL); + } + scales[ib] = scale; + + const float abs_scale = fabsf(scale); + if (abs_scale > max_abs_scale) { + max_abs_scale = abs_scale; + max_scale = scale; } + } - if (!amax) { - y[i].d = 0; - memset(y[i].qs, 0, QK_K); + + if (!max_abs_scale) { + memset(&y[i], 0, sizeof(block_q6_K)); + y[i].d = GGML_FP32_TO_FP16(0.f); x += QK_K; continue; } - const float iscale = -128.f/max; - for (int j = 0; j < QK_K; ++j) { - int v = nearest_int(iscale*x[j]); - y[i].qs[j] = MIN(127, v); + + float iscale = -128.f/max_scale; + y[i].d = GGML_FP32_TO_FP16(1/iscale); + for (int ib = 0; ib < QK_K/16; ++ib) { + y[i].scales[ib] = MIN(127, nearest_int(iscale*scales[ib])); } + for (int j = 0; j < QK_K/16; ++j) { - int sum = 0; + float d = GGML_FP16_TO_FP32(y[i].d) * y[i].scales[j]; + if (!d) { + continue; + } for (int ii = 0; ii < 16; ++ii) { - sum += y[i].qs[j*16 + ii]; + int l = nearest_int(x[16*j + ii]/d); + l = MAX(-32, MIN(31, l)); + L[16*j + ii] = l + 32; } - y[i].bsums[j] = sum; } - y[i].d = 1/iscale; + + uint8_t * restrict ql = y[i].ql; + uint8_t * restrict qh = y[i].qh; + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) { + const uint8_t q1 = L[j + l + 0] & 0xF; + const uint8_t q2 = L[j + l + 32] & 0xF; + const uint8_t q3 = L[j + l + 64] & 0xF; + const uint8_t q4 = L[j + l + 96] & 0xF; + ql[l+ 0] = q1 | (q3 << 4); + ql[l+32] = q2 | (q4 << 4); + qh[l] = (L[j + l] >> 4) | ((L[j + l + 32] >> 4) << 2) | ((L[j + l + 64] >> 4) << 4) | ((L[j + l + 96] >> 4) << 6); + } + ql += 64; + qh += 32; + } + x += QK_K; + } +#endif } -void dequantize_row_q8_K(const block_q8_K * restrict x, float * restrict y, int k) { - assert(k % QK_K == 0); - const int nb = k / QK_K; - - for (int i = 0; i < nb; i++) { - for (int j = 0; j < QK_K; ++j) { - *y++ = x[i].d * x[i].qs[j]; +size_t quantize_q6_K(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + size_t row_size = ggml_row_size(GGML_TYPE_Q6_K, n_per_row); + if (!quant_weights) { + quantize_row_q6_K_reference(src, dst, (int64_t)nrow*n_per_row); + } + else { + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q6_K_impl(src, (block_q6_K*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; } } + return nrow * row_size; } -void quantize_row_q8_K(const float * restrict x, void * restrict y, int k) { - quantize_row_q8_K_reference(x, y, k); -} +static void quantize_row_q4_0_impl(const float * restrict x, block_q4_0 * restrict y, int64_t n_per_row, const float * quant_weights) { + static_assert(QK4_0 == 32, "QK4_0 must be 32"); -//===================================== Dot ptoducts ================================= + if (!quant_weights) { + quantize_row_q4_0_reference(x, y, n_per_row); + return; + } -// -// Helper functions -// -#if __AVX__ || __AVX2__ || __AVX512F__ + float weight[QK4_0]; + int8_t L[QK4_0]; -// shuffles to pick the required scales in dot products -static inline __m256i get_scale_shuffle_q3k(int i) { - static const uint8_t k_shuffle[128] = { - 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, - 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, - 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, - 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15, - }; - return _mm256_loadu_si256((const __m256i*)k_shuffle + i); -} -static inline __m256i get_scale_shuffle_k4(int i) { - static const uint8_t k_shuffle[256] = { - 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, - 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, - 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, - 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, - 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, - 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, - 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, - 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15 - }; - return _mm256_loadu_si256((const __m256i*)k_shuffle + i); + float sum_x2 = 0; + for (int j = 0; j < n_per_row; ++j) sum_x2 += x[j]*x[j]; + float sigma2 = sum_x2/n_per_row; + + const int64_t nb = n_per_row/QK4_0; + for (int ib = 0; ib < nb; ++ib) { + const float * xb = x + QK4_0 * ib; + const float * qw = quant_weights + QK4_0 * ib; + for (int j = 0; j < QK4_0; ++j) weight[j] = qw[j] * sqrtf(sigma2 + xb[j]*xb[j]); + float d = make_qx_quants(QK4_0, 8, xb, L, 1, weight); + y[ib].d = GGML_FP32_TO_FP16(d); + for (int j = 0; j < 16; ++j) { + y[ib].qs[j] = L[j] | (L[j+16] << 4); + } + } } -static inline __m128i get_scale_shuffle(int i) { - static const uint8_t k_shuffle[128] = { - 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, - 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, - 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, - 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, - 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, - 10,10,10,10,10,10,10,10, 11,11,11,11,11,11,11,11, - 12,12,12,12,12,12,12,12, 13,13,13,13,13,13,13,13, - 14,14,14,14,14,14,14,14, 15,15,15,15,15,15,15,15 - }; - return _mm_loadu_si128((const __m128i*)k_shuffle + i); + +size_t quantize_q4_0(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + if (!quant_weights) { + quantize_row_q4_0_reference(src, dst, (int64_t)nrow*n_per_row); + return nrow * ggml_row_size(GGML_TYPE_Q4_0, n_per_row); + } + size_t row_size = ggml_row_size(GGML_TYPE_Q4_0, n_per_row); + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q4_0_impl(src, (block_q4_0*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + return nrow * row_size; } -#endif -void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int qk = QK8_0; - const int nb = n / qk; +static void quantize_row_q4_1_impl(const float * restrict x, block_q4_1 * restrict y, int64_t n_per_row, const float * quant_weights) { + static_assert(QK4_1 == 32, "QK4_1 must be 32"); - assert(n % qk == 0); + if (!quant_weights) { + quantize_row_q4_1_reference(x, y, n_per_row); + return; + } - const block_q4_0 * restrict x = vx; - const block_q8_0 * restrict y = vy; + float weight[QK4_1]; + uint8_t L[QK4_1], Laux[QK4_1]; -#if defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); + float sum_x2 = 0; + for (int j = 0; j < n_per_row; ++j) sum_x2 += x[j]*x[j]; + float sigma2 = sum_x2/n_per_row; - assert(nb % 2 == 0); // TODO: handle odd nb + const int64_t nb = n_per_row/QK4_1; + for (int ib = 0; ib < nb; ++ib) { + const float * xb = x + QK4_1 * ib; + const float * qw = quant_weights + QK4_1 * ib; + for (int j = 0; j < QK4_1; ++j) weight[j] = qw[j] * sqrtf(sigma2 + xb[j]*xb[j]); + float min; + float d = make_qkx3_quants(QK4_1, 15, xb, weight, L, &min, Laux, -0.9f, 0.05f, 36, false); + y[ib].d = GGML_FP32_TO_FP16(d); + y[ib].m = GGML_FP32_TO_FP16(-min); + for (int j = 0; j < 16; ++j) { + y[ib].qs[j] = L[j] | (L[j+16] << 4); + } + } +} - for (int i = 0; i < nb; i += 2) { - const block_q4_0 * restrict x0 = &x[i + 0]; - const block_q4_0 * restrict x1 = &x[i + 1]; - const block_q8_0 * restrict y0 = &y[i + 0]; - const block_q8_0 * restrict y1 = &y[i + 1]; +size_t quantize_q4_1(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + if (!quant_weights) { + quantize_row_q4_1_reference(src, dst, (int64_t)nrow*n_per_row); + return nrow * ggml_row_size(GGML_TYPE_Q4_1, n_per_row); + } + size_t row_size = ggml_row_size(GGML_TYPE_Q4_1, n_per_row); + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q4_1_impl(src, (block_q4_1*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + return nrow * row_size; +} - const uint8x16_t m4b = vdupq_n_u8(0x0F); - const int8x16_t s8b = vdupq_n_s8(0x8); +static void quantize_row_q5_0_impl(const float * restrict x, block_q5_0 * restrict y, int64_t n_per_row, const float * quant_weights) { + static_assert(QK5_0 == 32, "QK5_0 must be 32"); - const uint8x16_t v0_0 = vld1q_u8(x0->qs); - const uint8x16_t v0_1 = vld1q_u8(x1->qs); + if (!quant_weights) { + quantize_row_q5_0_reference(x, y, n_per_row); + return; + } - // 4-bit -> 8-bit - const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + float weight[QK5_0]; + int8_t L[QK5_0]; - // sub 8 - const int8x16_t v0_0ls = vsubq_s8(v0_0l, s8b); - const int8x16_t v0_0hs = vsubq_s8(v0_0h, s8b); - const int8x16_t v0_1ls = vsubq_s8(v0_1l, s8b); - const int8x16_t v0_1hs = vsubq_s8(v0_1h, s8b); + float sum_x2 = 0; + for (int j = 0; j < n_per_row; ++j) sum_x2 += x[j]*x[j]; + float sigma2 = sum_x2/n_per_row; - // load y - const int8x16_t v1_0l = vld1q_s8(y0->qs); - const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); - const int8x16_t v1_1l = vld1q_s8(y1->qs); - const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + const int64_t nb = n_per_row/QK5_0; + for (int ib = 0; ib < nb; ++ib) { + const float * xb = x + QK5_0 * ib; + const float * qw = quant_weights + QK5_0 * ib; + for (int j = 0; j < QK5_0; ++j) weight[j] = qw[j] * sqrtf(sigma2 + xb[j]*xb[j]); + float d = make_qx_quants(QK5_0, 16, xb, L, 1, weight); + y[ib].d = GGML_FP32_TO_FP16(d); -#if defined(__ARM_FEATURE_DOTPROD) - // dot product into int32x4_t - const int32x4_t p_0 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0l), v0_0hs, v1_0h); - const int32x4_t p_1 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1l), v0_1hs, v1_1h); + uint32_t qh = 0; - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); -#else - const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0ls), vget_low_s8 (v1_0l)); - const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0ls), vget_high_s8(v1_0l)); - const int16x8_t ph0l = vmull_s8(vget_low_s8 (v0_0hs), vget_low_s8 (v1_0h)); - const int16x8_t ph0h = vmull_s8(vget_high_s8(v0_0hs), vget_high_s8(v1_0h)); - - const int16x8_t pl1l = vmull_s8(vget_low_s8 (v0_1ls), vget_low_s8 (v1_1l)); - const int16x8_t pl1h = vmull_s8(vget_high_s8(v0_1ls), vget_high_s8(v1_1l)); - const int16x8_t ph1l = vmull_s8(vget_low_s8 (v0_1hs), vget_low_s8 (v1_1h)); - const int16x8_t ph1h = vmull_s8(vget_high_s8(v0_1hs), vget_high_s8(v1_1h)); - - const int32x4_t pl0 = vaddq_s32(vpaddlq_s16(pl0l), vpaddlq_s16(pl0h)); - const int32x4_t ph0 = vaddq_s32(vpaddlq_s16(ph0l), vpaddlq_s16(ph0h)); - const int32x4_t pl1 = vaddq_s32(vpaddlq_s16(pl1l), vpaddlq_s16(pl1h)); - const int32x4_t ph1 = vaddq_s32(vpaddlq_s16(ph1l), vpaddlq_s16(ph1h)); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32(pl0, ph0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32(pl1, ph1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); -#endif + for (int j = 0; j < 16; ++j) { + const uint8_t xi0 = L[j]; + const uint8_t xi1 = L[j+16]; + y[ib].qs[j] = (xi0 & 0x0F) | ((xi1 & 0x0F) << 4); + + // get the 5-th bit and store it in qh at the right position + qh |= ((xi0 & 0x10u) >> 4) << (j + 0); + qh |= ((xi1 & 0x10u) >> 4) << (j + QK5_0/2); + } + + memcpy(&y[ib].qh, &qh, sizeof(qh)); } +} - *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); -#elif defined(__AVX2__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); +size_t quantize_q5_0(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + if (!quant_weights) { + quantize_row_q5_0_reference(src, dst, (int64_t)nrow*n_per_row); + return nrow * ggml_row_size(GGML_TYPE_Q5_0, n_per_row); + } + size_t row_size = ggml_row_size(GGML_TYPE_Q5_0, n_per_row); + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q5_0_impl(src, (block_q5_0*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + return nrow * row_size; +} - // Main loop - for (int i = 0; i < nb; ++i) { - /* Compute combined scale for the block */ - const __m256 d = _mm256_set1_ps( GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d) ); +static void quantize_row_q5_1_impl(const float * restrict x, block_q5_1 * restrict y, int64_t n_per_row, const float * quant_weights) { + static_assert(QK5_1 == 32, "QK5_1 must be 32"); - __m256i bx = bytes_from_nibbles_32(x[i].qs); + if (!quant_weights) { + quantize_row_q5_1_reference(x, y, n_per_row); + return; + } - // Now we have a vector with bytes in [ 0 .. 15 ] interval. Offset them into [ -8 .. +7 ] interval. - const __m256i off = _mm256_set1_epi8( 8 ); - bx = _mm256_sub_epi8( bx, off ); + float weight[QK5_1]; + uint8_t L[QK5_1], Laux[QK5_1]; - __m256i by = _mm256_loadu_si256((const __m256i *)y[i].qs); + float sum_x2 = 0; + for (int j = 0; j < n_per_row; ++j) sum_x2 += x[j]*x[j]; + float sigma2 = sum_x2/n_per_row; - const __m256 q = mul_sum_i8_pairs_float(bx, by); + const int64_t nb = n_per_row/QK5_1; + for (int ib = 0; ib < nb; ++ib) { + const float * xb = x + QK5_1 * ib; + const float * qw = quant_weights + QK5_1 * ib; + for (int j = 0; j < QK5_1; ++j) weight[j] = qw[j] * sqrtf(sigma2 + xb[j]*xb[j]); + float min; + float d = make_qkx3_quants(QK5_1, 31, xb, weight, L, &min, Laux, -0.9f, 0.05f, 36, false); + y[ib].d = GGML_FP32_TO_FP16(d); + y[ib].m = GGML_FP32_TO_FP16(-min); - /* Multiply q with scale and accumulate */ - acc = _mm256_fmadd_ps( d, q, acc ); + uint32_t qh = 0; + for (int j = 0; j < 16; ++j) { + const uint8_t xi0 = L[j]; + const uint8_t xi1 = L[j+16]; + y[ib].qs[j] = (xi0 & 0x0F) | ((xi1 & 0x0F) << 4); + // get the 5-th bit and store it in qh at the right position + qh |= ((xi0 & 0x10u) >> 4) << (j + 0); + qh |= ((xi1 & 0x10u) >> 4) << (j + QK5_0/2); + } + memcpy(&y[ib].qh, &qh, sizeof(qh)); } +} - *s = hsum_float_8(acc); -#elif defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); +size_t quantize_q5_1(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + if (!quant_weights) { + quantize_row_q5_1_reference(src, dst, (int64_t)nrow*n_per_row); + return nrow * ggml_row_size(GGML_TYPE_Q5_1, n_per_row); + } + size_t row_size = ggml_row_size(GGML_TYPE_Q5_1, n_per_row); + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_q5_1_impl(src, (block_q5_1*)qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += row_size; + } + return nrow * row_size; +} - // Main loop - for (int i = 0; i < nb; ++i) { - // Compute combined scale for the block - const __m256 d = _mm256_set1_ps( GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d) ); +size_t quantize_q8_0(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + (void)quant_weights; // not used + const size_t row_size = ggml_row_size(GGML_TYPE_Q8_0, n_per_row); + quantize_row_q8_0_reference(src, dst, (int64_t)nrow*n_per_row); + return nrow * row_size; +} - const __m128i lowMask = _mm_set1_epi8(0xF); - const __m128i off = _mm_set1_epi8(8); +// ====================== "True" 2-bit (de)-quantization - const __m128i tmp = _mm_loadu_si128((const __m128i *)x[i].qs); +void dequantize_row_iq2_xxs(const block_iq2_xxs * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - __m128i bx = _mm_and_si128(lowMask, tmp); - __m128i by = _mm_loadu_si128((const __m128i *)y[i].qs); - bx = _mm_sub_epi8(bx, off); - const __m128i i32_0 = mul_sum_i8_pairs(bx, by); + uint32_t aux32[2]; + const uint8_t * aux8 = (const uint8_t *)aux32; - bx = _mm_and_si128(lowMask, _mm_srli_epi64(tmp, 4)); - by = _mm_loadu_si128((const __m128i *)(y[i].qs + 16)); - bx = _mm_sub_epi8(bx, off); - const __m128i i32_1 = mul_sum_i8_pairs(bx, by); + for (int i = 0; i < nb; i++) { - // Convert int32_t to float - __m256 p = _mm256_cvtepi32_ps(MM256_SET_M128I(i32_0, i32_1)); + const float d = GGML_FP16_TO_FP32(x[i].d); - // Apply the scale, and accumulate - acc = _mm256_add_ps(_mm256_mul_ps( d, p ), acc); + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + memcpy(aux32, x[i].qs + 4*ib32, 2*sizeof(uint32_t)); + const float db = d * (0.5f + (aux32[1] >> 28)) * 0.25f; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xxs_grid + aux8[l]); + const uint8_t signs = ksigns_iq2xs[(aux32[1] >> 7*l) & 127]; + for (int j = 0; j < 8; ++j) { + y[j] = db * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f); + } + y += 8; + } + } } +} - *s = hsum_float_8(acc); -#elif defined(__SSSE3__) - // set constants - const __m128i lowMask = _mm_set1_epi8(0xF); - const __m128i off = _mm_set1_epi8(8); - - // Initialize accumulator with zeros - __m128 acc_0 = _mm_setzero_ps(); - __m128 acc_1 = _mm_setzero_ps(); - __m128 acc_2 = _mm_setzero_ps(); - __m128 acc_3 = _mm_setzero_ps(); - - // First round without accumulation - { - _mm_prefetch(&x[0] + sizeof(block_q4_0), _MM_HINT_T0); - _mm_prefetch(&y[0] + sizeof(block_q8_0), _MM_HINT_T0); - - // Compute combined scale for the block 0 and 1 - const __m128 d_0_1 = _mm_set1_ps( GGML_FP16_TO_FP32(x[0].d) * GGML_FP16_TO_FP32(y[0].d) ); +// ====================== 2.3125 bpw (de)-quantization - const __m128i tmp_0_1 = _mm_loadu_si128((const __m128i *)x[0].qs); +void dequantize_row_iq2_xs(const block_iq2_xs * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - __m128i bx_0 = _mm_and_si128(lowMask, tmp_0_1); - __m128i by_0 = _mm_loadu_si128((const __m128i *)y[0].qs); - bx_0 = _mm_sub_epi8(bx_0, off); - const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); + float db[2]; - __m128i bx_1 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_0_1, 4)); - __m128i by_1 = _mm_loadu_si128((const __m128i *)(y[0].qs + 16)); - bx_1 = _mm_sub_epi8(bx_1, off); - const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); + for (int i = 0; i < nb; i++) { - _mm_prefetch(&x[1] + sizeof(block_q4_0), _MM_HINT_T0); - _mm_prefetch(&y[1] + sizeof(block_q8_0), _MM_HINT_T0); + const float d = GGML_FP16_TO_FP32(x[i].d); - // Compute combined scale for the block 2 and 3 - const __m128 d_2_3 = _mm_set1_ps( GGML_FP16_TO_FP32(x[1].d) * GGML_FP16_TO_FP32(y[1].d) ); + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + db[0] = d * (0.5f + (x[i].scales[ib32] & 0xf)) * 0.25f; + db[1] = d * (0.5f + (x[i].scales[ib32] >> 4)) * 0.25f; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (x[i].qs[4*ib32 + l] & 511)); + const uint8_t signs = ksigns_iq2xs[x[i].qs[4*ib32 + l] >> 9]; + for (int j = 0; j < 8; ++j) { + y[j] = db[l/2] * grid[j] * (signs & kmask_iq2xs[j] ? -1.f : 1.f); + } + y += 8; + } + } + } +} - const __m128i tmp_2_3 = _mm_loadu_si128((const __m128i *)x[1].qs); +// ====================== 2.5625 bpw (de)-quantization - __m128i bx_2 = _mm_and_si128(lowMask, tmp_2_3); - __m128i by_2 = _mm_loadu_si128((const __m128i *)y[1].qs); - bx_2 = _mm_sub_epi8(bx_2, off); - const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); +void dequantize_row_iq2_s(const block_iq2_s * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - __m128i bx_3 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_2_3, 4)); - __m128i by_3 = _mm_loadu_si128((const __m128i *)(y[1].qs + 16)); - bx_3 = _mm_sub_epi8(bx_3, off); - const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); + float db[2]; - // Convert int32_t to float - __m128 p0 = _mm_cvtepi32_ps(i32_0); - __m128 p1 = _mm_cvtepi32_ps(i32_1); - __m128 p2 = _mm_cvtepi32_ps(i32_2); - __m128 p3 = _mm_cvtepi32_ps(i32_3); + for (int i = 0; i < nb; i++) { - // Apply the scale - acc_0 = _mm_mul_ps( d_0_1, p0 ); - acc_1 = _mm_mul_ps( d_0_1, p1 ); - acc_2 = _mm_mul_ps( d_2_3, p2 ); - acc_3 = _mm_mul_ps( d_2_3, p3 ); + const float d = GGML_FP16_TO_FP32(x[i].d); + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint8_t * signs = qs + QK_K/8; + + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + db[0] = d * (0.5f + (x[i].scales[ib32] & 0xf)) * 0.25f; + db[1] = d * (0.5f + (x[i].scales[ib32] >> 4)) * 0.25f; + for (int l = 0; l < 4; ++l) { + const float dl = db[l/2]; + const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); + for (int j = 0; j < 8; ++j) { + y[j] = dl * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1.f : 1.f); + } + y += 8; + } + qs += 4; + signs += 4; + } } +} - assert(nb % 2 == 0); // TODO: handle odd nb - - // Main loop - for (int i = 2; i < nb; i+=2) { - _mm_prefetch(&x[i] + sizeof(block_q4_0), _MM_HINT_T0); - _mm_prefetch(&y[i] + sizeof(block_q8_0), _MM_HINT_T0); +// ====================== 3.0625 bpw (de)-quantization - // Compute combined scale for the block 0 and 1 - const __m128 d_0_1 = _mm_set1_ps( GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d) ); +void dequantize_row_iq3_xxs(const block_iq3_xxs * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - const __m128i tmp_0_1 = _mm_loadu_si128((const __m128i *)x[i].qs); + uint32_t aux32; - __m128i bx_0 = _mm_and_si128(lowMask, tmp_0_1); - __m128i by_0 = _mm_loadu_si128((const __m128i *)y[i].qs); - bx_0 = _mm_sub_epi8(bx_0, off); - const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); + for (int i = 0; i < nb; i++) { - __m128i bx_1 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_0_1, 4)); - __m128i by_1 = _mm_loadu_si128((const __m128i *)(y[i].qs + 16)); - bx_1 = _mm_sub_epi8(bx_1, off); - const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); + const float d = GGML_FP16_TO_FP32(x[i].d); + const uint8_t * qs = x[i].qs; + const uint8_t * scales_and_signs = qs + QK_K/4; + + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + memcpy(&aux32, scales_and_signs + 4*ib32, sizeof(uint32_t)); + const float db = d * (0.5f + (aux32 >> 28)) * 0.5f; + for (int l = 0; l < 4; ++l) { + const uint8_t signs = ksigns_iq2xs[(aux32 >> 7*l) & 127]; + const uint8_t * grid1 = (const uint8_t *)(iq3xxs_grid + qs[2*l+0]); + const uint8_t * grid2 = (const uint8_t *)(iq3xxs_grid + qs[2*l+1]); + for (int j = 0; j < 4; ++j) { + y[j+0] = db * grid1[j] * (signs & kmask_iq2xs[j+0] ? -1.f : 1.f); + y[j+4] = db * grid2[j] * (signs & kmask_iq2xs[j+4] ? -1.f : 1.f); + } + y += 8; + } + qs += 8; + } + } +} - _mm_prefetch(&x[i] + 2 * sizeof(block_q4_0), _MM_HINT_T0); - _mm_prefetch(&y[i] + 2 * sizeof(block_q8_0), _MM_HINT_T0); +// ====================== 3.3125 bpw (de)-quantization - // Compute combined scale for the block 2 and 3 - const __m128 d_2_3 = _mm_set1_ps( GGML_FP16_TO_FP32(x[i + 1].d) * GGML_FP16_TO_FP32(y[i + 1].d) ); +void dequantize_row_iq3_s(const block_iq3_s * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - const __m128i tmp_2_3 = _mm_loadu_si128((const __m128i *)x[i + 1].qs); + for (int i = 0; i < nb; i++) { - __m128i bx_2 = _mm_and_si128(lowMask, tmp_2_3); - __m128i by_2 = _mm_loadu_si128((const __m128i *)y[i + 1].qs); - bx_2 = _mm_sub_epi8(bx_2, off); - const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); + const float d = GGML_FP16_TO_FP32(x[i].d); + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint8_t * signs = x[i].signs; + + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const float db1 = d * (1 + 2*(x[i].scales[ib32/2] & 0xf)); + const float db2 = d * (1 + 2*(x[i].scales[ib32/2] >> 4)); + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[0] << (8-2*l)) & 256))); + const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[0] << (7-2*l)) & 256))); + for (int j = 0; j < 4; ++j) { + y[j+0] = db1 * grid1[j] * (signs[l] & kmask_iq2xs[j+0] ? -1.f : 1.f); + y[j+4] = db1 * grid2[j] * (signs[l] & kmask_iq2xs[j+4] ? -1.f : 1.f); + } + y += 8; + } + qs += 8; + signs += 4; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[1] << (8-2*l)) & 256))); + const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[1] << (7-2*l)) & 256))); + for (int j = 0; j < 4; ++j) { + y[j+0] = db2 * grid1[j] * (signs[l] & kmask_iq2xs[j+0] ? -1.f : 1.f); + y[j+4] = db2 * grid2[j] * (signs[l] & kmask_iq2xs[j+4] ? -1.f : 1.f); + } + y += 8; + } + qh += 2; + qs += 8; + signs += 4; + } + } +} - __m128i bx_3 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_2_3, 4)); - __m128i by_3 = _mm_loadu_si128((const __m128i *)(y[i + 1].qs + 16)); - bx_3 = _mm_sub_epi8(bx_3, off); - const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); +// ====================== 1.5625 bpw (de)-quantization - // Convert int32_t to float - __m128 p0 = _mm_cvtepi32_ps(i32_0); - __m128 p1 = _mm_cvtepi32_ps(i32_1); - __m128 p2 = _mm_cvtepi32_ps(i32_2); - __m128 p3 = _mm_cvtepi32_ps(i32_3); +void dequantize_row_iq1_s(const block_iq1_s * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - // Apply the scale - __m128 p0_d = _mm_mul_ps( d_0_1, p0 ); - __m128 p1_d = _mm_mul_ps( d_0_1, p1 ); - __m128 p2_d = _mm_mul_ps( d_2_3, p2 ); - __m128 p3_d = _mm_mul_ps( d_2_3, p3 ); + for (int i = 0; i < nb; i++) { - // Acummulate - acc_0 = _mm_add_ps(p0_d, acc_0); - acc_1 = _mm_add_ps(p1_d, acc_1); - acc_2 = _mm_add_ps(p2_d, acc_2); - acc_3 = _mm_add_ps(p3_d, acc_3); + const float d = GGML_FP16_TO_FP32(x[i].d); + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + for (int ib = 0; ib < QK_K/32; ++ib) { + const float dl = d * (2*((qh[ib] >> 12) & 7) + 1); + const float delta = qh[ib] & 0x8000 ? -IQ1S_DELTA : IQ1S_DELTA; + for (int l = 0; l < 4; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((qh[ib] >> 3*l) & 7) << 8))); + for (int j = 0; j < 8; ++j) { + y[j] = dl * (grid[j] + delta); + } + y += 8; + } + qs += 4; + } } +} - *s = hsum_float_4x4(acc_0, acc_1, acc_2, acc_3); -#elif defined(__riscv_v_intrinsic) - float sumf = 0.0; - - size_t vl = __riscv_vsetvl_e8m1(qk/2); +void dequantize_row_iq1_m(const block_iq1_m * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; - for (int i = 0; i < nb; i++) { - // load elements - vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[i].qs, vl); + float delta[4]; + uint16_t idx[4]; - vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[i].qs, vl); - vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[i].qs+16, vl); +#if QK_K != 64 + iq1m_scale_t scale; +#endif - // mask and store lower part of x, and then upper part - vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); + for (int i = 0; i < nb; i++) { - vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + const uint16_t * sc = (const uint16_t *)x[i].scales; +#if QK_K == 64 + const float d = GGML_FP16_TO_FP32(x[i].d); +#else + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); + const float d = GGML_FP16_TO_FP32(scale.f16); +#endif + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; - // subtract offset - vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 8, vl); - vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 8, vl); + for (int ib = 0; ib < QK_K/32; ++ib) { +#if QK_K == 64 + const float dl1 = d * (2*((sc[ib/2] >> (8*(ib%2)+0)) & 0xf) + 1); + const float dl2 = d * (2*((sc[ib/2] >> (8*(ib%2)+4)) & 0xf) + 1); +#else + const float dl1 = d * (2*((sc[ib/2] >> (6*(ib%2)+0)) & 0x7) + 1); + const float dl2 = d * (2*((sc[ib/2] >> (6*(ib%2)+3)) & 0x7) + 1); +#endif + idx[0] = qs[0] | ((qh[0] << 8) & 0x700); + idx[1] = qs[1] | ((qh[0] << 4) & 0x700); + idx[2] = qs[2] | ((qh[1] << 8) & 0x700); + idx[3] = qs[3] | ((qh[1] << 4) & 0x700); + delta[0] = qh[0] & 0x08 ? -IQ1S_DELTA : IQ1S_DELTA; + delta[1] = qh[0] & 0x80 ? -IQ1S_DELTA : IQ1S_DELTA; + delta[2] = qh[1] & 0x08 ? -IQ1S_DELTA : IQ1S_DELTA; + delta[3] = qh[1] & 0x80 ? -IQ1S_DELTA : IQ1S_DELTA; + for (int l = 0; l < 2; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + idx[l]); + for (int j = 0; j < 8; ++j) { + y[j] = dl1 * (grid[j] + delta[l]); + } + y += 8; + } + for (int l = 2; l < 4; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + idx[l]); + for (int j = 0; j < 8; ++j) { + y[j] = dl2 * (grid[j] + delta[l]); + } + y += 8; + } + qs += 4; + qh += 2; + } + } +} - vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); - vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); +static const int8_t kvalues_iq4nl[16] = {-127, -104, -83, -65, -49, -35, -22, -10, 1, 13, 25, 38, 53, 69, 89, 113}; - vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); +void dequantize_row_iq4_nl(const block_iq4_nl * restrict x, float * restrict y, int64_t k) { + assert(k % QK4_NL == 0); + const int64_t nb = k / QK4_NL; - vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); - vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); + for (int i = 0; i < nb; i++) { - int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); + const uint8_t * qs = x[i].qs; - sumf += sumi*GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d); + const float d = GGML_FP16_TO_FP32(x[i].d); + for (int j = 0; j < QK4_NL/2; ++j) { + y[j+ 0] = d * kvalues_iq4nl[qs[j] & 0xf]; + y[j+QK4_NL/2] = d * kvalues_iq4nl[qs[j] >> 4]; + } + y += QK4_NL; + qs += QK4_NL/2; } +} - *s = sumf; +void dequantize_row_iq4_xs(const block_iq4_xs * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); +#if QK_K == 64 + dequantize_row_iq4_nl((const block_iq4_nl *)x, y, k); #else - // scalar - float sumf = 0.0; + const int64_t nb = k / QK_K; for (int i = 0; i < nb; i++) { - int sumi = 0; - for (int j = 0; j < qk/2; ++j) { - const int v0 = (x[i].qs[j] & 0x0F) - 8; - const int v1 = (x[i].qs[j] >> 4) - 8; + const uint8_t * qs = x[i].qs; - sumi += (v0 * y[i].qs[j]) + (v1 * y[i].qs[j + qk/2]); + const float d = GGML_FP16_TO_FP32(x[i].d); + + for (int ib = 0; ib < QK_K/32; ++ib) { + const int ls = ((x[i].scales_l[ib/2] >> 4*(ib%2)) & 0xf) | (((x[i].scales_h >> 2*ib) & 3) << 4); + const float dl = d * (ls - 32); + for (int j = 0; j < 16; ++j) { + y[j+ 0] = dl * kvalues_iq4nl[qs[j] & 0xf]; + y[j+16] = dl * kvalues_iq4nl[qs[j] >> 4]; + } + y += 32; + qs += 16; } + } +#endif +} - sumf += sumi*GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d); +//===================================== Q8_K ============================================== + +void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; + + for (int i = 0; i < nb; i++) { + + float max = 0; + float amax = 0; + for (int j = 0; j < QK_K; ++j) { + float ax = fabsf(x[j]); + if (ax > amax) { + amax = ax; max = x[j]; + } + } + if (!amax) { + y[i].d = 0; + memset(y[i].qs, 0, QK_K); + x += QK_K; + continue; + } + //const float iscale = -128.f/max; + // We need this change for IQ2_XXS, else the AVX implementation becomes very awkward + const float iscale = -127.f/max; + for (int j = 0; j < QK_K; ++j) { + int v = nearest_int(iscale*x[j]); + y[i].qs[j] = MIN(127, v); + } + for (int j = 0; j < QK_K/16; ++j) { + int sum = 0; + for (int ii = 0; ii < 16; ++ii) { + sum += y[i].qs[j*16 + ii]; + } + y[i].bsums[j] = sum; + } + y[i].d = 1/iscale; + x += QK_K; } +} - *s = sumf; -#endif +void dequantize_row_q8_K(const block_q8_K * restrict x, float * restrict y, int64_t k) { + assert(k % QK_K == 0); + const int64_t nb = k / QK_K; + + for (int i = 0; i < nb; i++) { + for (int j = 0; j < QK_K; ++j) { + *y++ = x[i].d * x[i].qs[j]; + } + } } -void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int qk = QK8_1; +void quantize_row_q8_K(const float * restrict x, void * restrict y, int64_t k) { + quantize_row_q8_K_reference(x, y, k); +} + +//===================================== Dot ptoducts ================================= + +// +// Helper functions +// +#if __AVX__ || __AVX2__ || __AVX512F__ + +// shuffles to pick the required scales in dot products +static inline __m256i get_scale_shuffle_q3k(int i) { + static const uint8_t k_shuffle[128] = { + 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, + 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, + 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, + 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15, + }; + return _mm256_loadu_si256((const __m256i*)k_shuffle + i); +} +static inline __m256i get_scale_shuffle_k4(int i) { + static const uint8_t k_shuffle[256] = { + 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1, + 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, 2, 3, + 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, 4, 5, + 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, 6, 7, + 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, 8, 9, + 10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11,10,11, + 12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13,12,13, + 14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15,14,15 + }; + return _mm256_loadu_si256((const __m256i*)k_shuffle + i); +} +static inline __m128i get_scale_shuffle(int i) { + static const uint8_t k_shuffle[128] = { + 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, + 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, + 4, 4, 4, 4, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 5, + 6, 6, 6, 6, 6, 6, 6, 6, 7, 7, 7, 7, 7, 7, 7, 7, + 8, 8, 8, 8, 8, 8, 8, 8, 9, 9, 9, 9, 9, 9, 9, 9, + 10,10,10,10,10,10,10,10, 11,11,11,11,11,11,11,11, + 12,12,12,12,12,12,12,12, 13,13,13,13,13,13,13,13, + 14,14,14,14,14,14,14,14, 15,15,15,15,15,15,15,15 + }; + return _mm_loadu_si128((const __m128i*)k_shuffle + i); +} +#endif + +void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_0; const int nb = n / qk; assert(n % qk == 0); +#if defined(__ARM_FEATURE_MATMUL_INT8) + assert((nrc == 2) || (nrc == 1)); +#else + assert(nrc == 1); +#endif + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q4_1 * restrict x = vx; - const block_q8_1 * restrict y = vy; + const block_q4_0 * restrict x = vx; + const block_q8_0 * restrict y = vy; - // TODO: add WASM SIMD +#if defined(__ARM_FEATURE_MATMUL_INT8) + if (nrc == 2) { + const block_q4_0 * restrict vx0 = vx; + const block_q4_0 * restrict vx1 = vx + bx; + + const block_q8_0 * restrict vy0 = vy; + const block_q8_0 * restrict vy1 = vy + by; + + float32x4_t sumv0 = vdupq_n_f32(0.0f); + + for (int i = 0; i < nb; i++) { + const block_q4_0 * restrict b_x0 = &vx0[i]; + const block_q4_0 * restrict b_x1 = &vx1[i]; + const block_q8_0 * restrict b_y0 = &vy0[i]; + const block_q8_0 * restrict b_y1 = &vy1[i]; + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + const int8x16_t s8b = vdupq_n_s8(0x8); + + const uint8x16_t v0_0 = vld1q_u8(b_x0->qs); + const uint8x16_t v0_1 = vld1q_u8(b_x1->qs); + + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // sub 8 + const int8x16_t x0_l = vsubq_s8(v0_0l, s8b); + const int8x16_t x0_h = vsubq_s8(v0_0h, s8b); + const int8x16_t x1_l = vsubq_s8(v0_1l, s8b); + const int8x16_t x1_h = vsubq_s8(v0_1h, s8b); + + // load y + const int8x16_t y0_l = vld1q_s8(b_y0->qs); + const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); + const int8x16_t y1_l = vld1q_s8(b_y1->qs); + const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); + + float32x4_t scale = {GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d)}; + + int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + + int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + + int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + + int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + + sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), + l1, r1)), l2, r2)), l3, r3))), scale); + } + float32x4_t sumv1 = vextq_f32(sumv0, sumv0, 2); + float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); + + vst1_f32(s, vget_low_f32(sumv2)); + vst1_f32(s + bs, vget_high_f32(sumv2)); + return; + } +#endif #if defined(__ARM_NEON) float32x4_t sumv0 = vdupq_n_f32(0.0f); float32x4_t sumv1 = vdupq_n_f32(0.0f); - float summs = 0; - assert(nb % 2 == 0); // TODO: handle odd nb for (int i = 0; i < nb; i += 2) { - const block_q4_1 * restrict x0 = &x[i + 0]; - const block_q4_1 * restrict x1 = &x[i + 1]; - const block_q8_1 * restrict y0 = &y[i + 0]; - const block_q8_1 * restrict y1 = &y[i + 1]; - - summs += GGML_FP16_TO_FP32(x0->m) * y0->s + GGML_FP16_TO_FP32(x1->m) * y1->s; + const block_q4_0 * restrict x0 = &x[i + 0]; + const block_q4_0 * restrict x1 = &x[i + 1]; + const block_q8_0 * restrict y0 = &y[i + 0]; + const block_q8_0 * restrict y1 = &y[i + 1]; const uint8x16_t m4b = vdupq_n_u8(0x0F); + const int8x16_t s8b = vdupq_n_s8(0x8); const uint8x16_t v0_0 = vld1q_u8(x0->qs); const uint8x16_t v0_1 = vld1q_u8(x1->qs); @@ -2670,393 +3791,445 @@ void ggml_vec_dot_q4_1_q8_1(const int n, float * restrict s, const void * restri const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + // sub 8 + const int8x16_t v0_0ls = vsubq_s8(v0_0l, s8b); + const int8x16_t v0_0hs = vsubq_s8(v0_0h, s8b); + const int8x16_t v0_1ls = vsubq_s8(v0_1l, s8b); + const int8x16_t v0_1hs = vsubq_s8(v0_1h, s8b); + // load y const int8x16_t v1_0l = vld1q_s8(y0->qs); const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); const int8x16_t v1_1l = vld1q_s8(y1->qs); const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); -#if defined(__ARM_FEATURE_DOTPROD) // dot product into int32x4_t - const int32x4_t p_0 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_0l, v1_0l), v0_0h, v1_0h); - const int32x4_t p_1 = vdotq_s32(vdotq_s32(vdupq_n_s32(0), v0_1l, v1_1l), v0_1h, v1_1h); + const int32x4_t p_0 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_0ls, v1_0l), v0_0hs, v1_0h); + const int32x4_t p_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_1ls, v1_1l), v0_1hs, v1_1h); - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*y0->d); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*y1->d); -#else - const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0l), vget_low_s8 (v1_0l)); - const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0l), vget_high_s8(v1_0l)); - const int16x8_t ph0l = vmull_s8(vget_low_s8 (v0_0h), vget_low_s8 (v1_0h)); - const int16x8_t ph0h = vmull_s8(vget_high_s8(v0_0h), vget_high_s8(v1_0h)); - - const int16x8_t pl1l = vmull_s8(vget_low_s8 (v0_1l), vget_low_s8 (v1_1l)); - const int16x8_t pl1h = vmull_s8(vget_high_s8(v0_1l), vget_high_s8(v1_1l)); - const int16x8_t ph1l = vmull_s8(vget_low_s8 (v0_1h), vget_low_s8 (v1_1h)); - const int16x8_t ph1h = vmull_s8(vget_high_s8(v0_1h), vget_high_s8(v1_1h)); - - const int32x4_t pl0 = vaddq_s32(vpaddlq_s16(pl0l), vpaddlq_s16(pl0h)); - const int32x4_t ph0 = vaddq_s32(vpaddlq_s16(ph0l), vpaddlq_s16(ph0h)); - const int32x4_t pl1 = vaddq_s32(vpaddlq_s16(pl1l), vpaddlq_s16(pl1h)); - const int32x4_t ph1 = vaddq_s32(vpaddlq_s16(ph1l), vpaddlq_s16(ph1h)); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32(pl0, ph0)), GGML_FP16_TO_FP32(x0->d)*y0->d); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32(pl1, ph1)), GGML_FP16_TO_FP32(x1->d)*y1->d); -#endif + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); } - *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs; -#elif defined(__AVX2__) || defined(__AVX__) + *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); +#elif defined(__AVX2__) // Initialize accumulator with zeros __m256 acc = _mm256_setzero_ps(); - float summs = 0; - // Main loop for (int i = 0; i < nb; ++i) { - const float d0 = GGML_FP16_TO_FP32(x[i].d); - const float d1 = y[i].d; - - summs += GGML_FP16_TO_FP32(x[i].m) * y[i].s; + /* Compute combined scale for the block */ + const __m256 d = _mm256_set1_ps( GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d) ); - const __m256 d0v = _mm256_set1_ps( d0 ); - const __m256 d1v = _mm256_set1_ps( d1 ); + __m256i qx = bytes_from_nibbles_32(x[i].qs); - // Compute combined scales - const __m256 d0d1 = _mm256_mul_ps( d0v, d1v ); + // Now we have a vector with bytes in [ 0 .. 15 ] interval. Offset them into [ -8 .. +7 ] interval. + const __m256i off = _mm256_set1_epi8( 8 ); + qx = _mm256_sub_epi8( qx, off ); - // Load 16 bytes, and unpack 4 bit fields into bytes, making 32 bytes - const __m256i bx = bytes_from_nibbles_32(x[i].qs); - const __m256i by = _mm256_loadu_si256( (const __m256i *)y[i].qs ); + __m256i qy = _mm256_loadu_si256((const __m256i *)y[i].qs); - const __m256 xy = mul_sum_us8_pairs_float(bx, by); + const __m256 q = mul_sum_i8_pairs_float(qx, qy); - // Accumulate d0*d1*x*y -#if defined(__AVX2__) - acc = _mm256_fmadd_ps( d0d1, xy, acc ); -#else - acc = _mm256_add_ps( _mm256_mul_ps( d0d1, xy ), acc ); -#endif + /* Multiply q with scale and accumulate */ + acc = _mm256_fmadd_ps( d, q, acc ); } - *s = hsum_float_8(acc) + summs; -#elif defined(__riscv_v_intrinsic) - float sumf = 0.0; + *s = hsum_float_8(acc); +#elif defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); - size_t vl = __riscv_vsetvl_e8m1(qk/2); + // Main loop + for (int i = 0; i < nb; ++i) { + // Compute combined scale for the block + const __m256 d = _mm256_set1_ps( GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d) ); - for (int i = 0; i < nb; i++) { - // load elements - vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[i].qs, vl); + const __m128i lowMask = _mm_set1_epi8(0xF); + const __m128i off = _mm_set1_epi8(8); - vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[i].qs, vl); - vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[i].qs+16, vl); + const __m128i tmp = _mm_loadu_si128((const __m128i *)x[i].qs); - // mask and store lower part of x, and then upper part - vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); + __m128i bx_0 = _mm_and_si128(lowMask, tmp); + __m128i by_0 = _mm_loadu_si128((const __m128i *)y[i].qs); + bx_0 = _mm_sub_epi8(bx_0, off); + const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); - vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + bx_0 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp, 4)); + by_0 = _mm_loadu_si128((const __m128i *)(y[i].qs + 16)); + bx_0 = _mm_sub_epi8(bx_0, off); + const __m128i i32_1 = mul_sum_i8_pairs(bx_0, by_0); - vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); - vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); + // Convert int32_t to float + __m256 p = _mm256_cvtepi32_ps(MM256_SET_M128I(i32_0, i32_1)); - vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); + // Apply the scale, and accumulate + acc = _mm256_add_ps(_mm256_mul_ps( d, p ), acc); + } - vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); - vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); + *s = hsum_float_8(acc); +#elif defined(__SSSE3__) + // set constants + const __m128i lowMask = _mm_set1_epi8(0xF); + const __m128i off = _mm_set1_epi8(8); - int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); + // Initialize accumulator with zeros + __m128 acc_0 = _mm_setzero_ps(); + __m128 acc_1 = _mm_setzero_ps(); + __m128 acc_2 = _mm_setzero_ps(); + __m128 acc_3 = _mm_setzero_ps(); - sumf += (GGML_FP16_TO_FP32(x[i].d)*y[i].d)*sumi + GGML_FP16_TO_FP32(x[i].m)*y[i].s; - } + // First round without accumulation + { + _mm_prefetch(&x[0] + sizeof(block_q4_0), _MM_HINT_T0); + _mm_prefetch(&y[0] + sizeof(block_q8_0), _MM_HINT_T0); - *s = sumf; -#else - // scalar - float sumf = 0.0; + // Compute combined scale for the block 0 and 1 + const __m128 d_0_1 = _mm_set1_ps( GGML_FP16_TO_FP32(x[0].d) * GGML_FP16_TO_FP32(y[0].d) ); - for (int i = 0; i < nb; i++) { - int sumi = 0; + const __m128i tmp_0_1 = _mm_loadu_si128((const __m128i *)x[0].qs); - for (int j = 0; j < qk/2; ++j) { - const int v0 = (x[i].qs[j] & 0x0F); - const int v1 = (x[i].qs[j] >> 4); + __m128i bx_0 = _mm_and_si128(lowMask, tmp_0_1); + __m128i by_0 = _mm_loadu_si128((const __m128i *)y[0].qs); + bx_0 = _mm_sub_epi8(bx_0, off); + const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); - sumi += (v0 * y[i].qs[j]) + (v1 * y[i].qs[j + qk/2]); - } + __m128i bx_1 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_0_1, 4)); + __m128i by_1 = _mm_loadu_si128((const __m128i *)(y[0].qs + 16)); + bx_1 = _mm_sub_epi8(bx_1, off); + const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); - sumf += (GGML_FP16_TO_FP32(x[i].d)*y[i].d)*sumi + GGML_FP16_TO_FP32(x[i].m)*y[i].s; - } + _mm_prefetch(&x[1] + sizeof(block_q4_0), _MM_HINT_T0); + _mm_prefetch(&y[1] + sizeof(block_q8_0), _MM_HINT_T0); - *s = sumf; -#endif -} - -void ggml_vec_dot_q5_0_q8_0(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int qk = QK8_0; - const int nb = n / qk; + // Compute combined scale for the block 2 and 3 + const __m128 d_2_3 = _mm_set1_ps( GGML_FP16_TO_FP32(x[1].d) * GGML_FP16_TO_FP32(y[1].d) ); - assert(n % qk == 0); - assert(qk == QK5_0); + const __m128i tmp_2_3 = _mm_loadu_si128((const __m128i *)x[1].qs); - const block_q5_0 * restrict x = vx; - const block_q8_0 * restrict y = vy; + __m128i bx_2 = _mm_and_si128(lowMask, tmp_2_3); + __m128i by_2 = _mm_loadu_si128((const __m128i *)y[1].qs); + bx_2 = _mm_sub_epi8(bx_2, off); + const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); -#if defined(__ARM_NEON) - float32x4_t sumv0 = vdupq_n_f32(0.0f); - float32x4_t sumv1 = vdupq_n_f32(0.0f); + __m128i bx_3 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_2_3, 4)); + __m128i by_3 = _mm_loadu_si128((const __m128i *)(y[1].qs + 16)); + bx_3 = _mm_sub_epi8(bx_3, off); + const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); - uint32_t qh0; - uint32_t qh1; + // Convert int32_t to float + __m128 p0 = _mm_cvtepi32_ps(i32_0); + __m128 p1 = _mm_cvtepi32_ps(i32_1); + __m128 p2 = _mm_cvtepi32_ps(i32_2); + __m128 p3 = _mm_cvtepi32_ps(i32_3); - uint64_t tmp0[4]; - uint64_t tmp1[4]; + // Apply the scale + acc_0 = _mm_mul_ps( d_0_1, p0 ); + acc_1 = _mm_mul_ps( d_0_1, p1 ); + acc_2 = _mm_mul_ps( d_2_3, p2 ); + acc_3 = _mm_mul_ps( d_2_3, p3 ); + } assert(nb % 2 == 0); // TODO: handle odd nb - for (int i = 0; i < nb; i += 2) { - const block_q5_0 * restrict x0 = &x[i]; - const block_q5_0 * restrict x1 = &x[i + 1]; - const block_q8_0 * restrict y0 = &y[i]; - const block_q8_0 * restrict y1 = &y[i + 1]; + // Main loop + for (int i = 2; i < nb; i+=2) { + _mm_prefetch(&x[i] + sizeof(block_q4_0), _MM_HINT_T0); + _mm_prefetch(&y[i] + sizeof(block_q8_0), _MM_HINT_T0); - const uint8x16_t m4b = vdupq_n_u8(0x0F); + // Compute combined scale for the block 0 and 1 + const __m128 d_0_1 = _mm_set1_ps( GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d) ); - // extract the 5th bit via lookup table ((!b) << 4) - memcpy(&qh0, x0->qh, sizeof(qh0)); - memcpy(&qh1, x1->qh, sizeof(qh1)); + const __m128i tmp_0_1 = _mm_loadu_si128((const __m128i *)x[i].qs); - tmp0[0] = table_b2b_1[(qh0 >> 0) & 0xFF]; - tmp0[1] = table_b2b_1[(qh0 >> 8) & 0xFF]; - tmp0[2] = table_b2b_1[(qh0 >> 16) & 0xFF]; - tmp0[3] = table_b2b_1[(qh0 >> 24) ]; + __m128i bx_0 = _mm_and_si128(lowMask, tmp_0_1); + __m128i by_0 = _mm_loadu_si128((const __m128i *)y[i].qs); + bx_0 = _mm_sub_epi8(bx_0, off); + const __m128i i32_0 = mul_sum_i8_pairs(bx_0, by_0); - tmp1[0] = table_b2b_1[(qh1 >> 0) & 0xFF]; - tmp1[1] = table_b2b_1[(qh1 >> 8) & 0xFF]; - tmp1[2] = table_b2b_1[(qh1 >> 16) & 0xFF]; - tmp1[3] = table_b2b_1[(qh1 >> 24) ]; + __m128i bx_1 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_0_1, 4)); + __m128i by_1 = _mm_loadu_si128((const __m128i *)(y[i].qs + 16)); + bx_1 = _mm_sub_epi8(bx_1, off); + const __m128i i32_1 = mul_sum_i8_pairs(bx_1, by_1); - const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); - const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); - const int8x16_t qhl1 = vld1q_s8((const int8_t *)(tmp1 + 0)); - const int8x16_t qhh1 = vld1q_s8((const int8_t *)(tmp1 + 2)); + _mm_prefetch(&x[i] + 2 * sizeof(block_q4_0), _MM_HINT_T0); + _mm_prefetch(&y[i] + 2 * sizeof(block_q8_0), _MM_HINT_T0); - const uint8x16_t v0_0 = vld1q_u8(x0->qs); - const uint8x16_t v0_1 = vld1q_u8(x1->qs); + // Compute combined scale for the block 2 and 3 + const __m128 d_2_3 = _mm_set1_ps( GGML_FP16_TO_FP32(x[i + 1].d) * GGML_FP16_TO_FP32(y[i + 1].d) ); - // 4-bit -> 8-bit - int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + const __m128i tmp_2_3 = _mm_loadu_si128((const __m128i *)x[i + 1].qs); - // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) - const int8x16_t v0_0lf = vsubq_s8(v0_0l, qhl0); - const int8x16_t v0_0hf = vsubq_s8(v0_0h, qhh0); - const int8x16_t v0_1lf = vsubq_s8(v0_1l, qhl1); - const int8x16_t v0_1hf = vsubq_s8(v0_1h, qhh1); + __m128i bx_2 = _mm_and_si128(lowMask, tmp_2_3); + __m128i by_2 = _mm_loadu_si128((const __m128i *)y[i + 1].qs); + bx_2 = _mm_sub_epi8(bx_2, off); + const __m128i i32_2 = mul_sum_i8_pairs(bx_2, by_2); - // load y - const int8x16_t v1_0l = vld1q_s8(y0->qs); - const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); - const int8x16_t v1_1l = vld1q_s8(y1->qs); - const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + __m128i bx_3 = _mm_and_si128(lowMask, _mm_srli_epi64(tmp_2_3, 4)); + __m128i by_3 = _mm_loadu_si128((const __m128i *)(y[i + 1].qs + 16)); + bx_3 = _mm_sub_epi8(bx_3, off); + const __m128i i32_3 = mul_sum_i8_pairs(bx_3, by_3); -#if defined(__ARM_FEATURE_DOTPROD) - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( - vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), - vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( - vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), - vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); -#else - const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0lf), vget_low_s8 (v1_0l)); - const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0lf), vget_high_s8(v1_0l)); - const int16x8_t ph0l = vmull_s8(vget_low_s8 (v0_0hf), vget_low_s8 (v1_0h)); - const int16x8_t ph0h = vmull_s8(vget_high_s8(v0_0hf), vget_high_s8(v1_0h)); - - const int16x8_t pl1l = vmull_s8(vget_low_s8 (v0_1lf), vget_low_s8 (v1_1l)); - const int16x8_t pl1h = vmull_s8(vget_high_s8(v0_1lf), vget_high_s8(v1_1l)); - const int16x8_t ph1l = vmull_s8(vget_low_s8 (v0_1hf), vget_low_s8 (v1_1h)); - const int16x8_t ph1h = vmull_s8(vget_high_s8(v0_1hf), vget_high_s8(v1_1h)); - - const int32x4_t pl0 = vaddq_s32(vpaddlq_s16(pl0l), vpaddlq_s16(pl0h)); - const int32x4_t ph0 = vaddq_s32(vpaddlq_s16(ph0l), vpaddlq_s16(ph0h)); - const int32x4_t pl1 = vaddq_s32(vpaddlq_s16(pl1l), vpaddlq_s16(pl1h)); - const int32x4_t ph1 = vaddq_s32(vpaddlq_s16(ph1l), vpaddlq_s16(ph1h)); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32(pl0, ph0)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32(pl1, ph1)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); -#endif - } + // Convert int32_t to float + __m128 p0 = _mm_cvtepi32_ps(i32_0); + __m128 p1 = _mm_cvtepi32_ps(i32_1); + __m128 p2 = _mm_cvtepi32_ps(i32_2); + __m128 p3 = _mm_cvtepi32_ps(i32_3); - *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); -#elif defined(__wasm_simd128__) - v128_t sumv = wasm_f32x4_splat(0.0f); + // Apply the scale + __m128 p0_d = _mm_mul_ps( d_0_1, p0 ); + __m128 p1_d = _mm_mul_ps( d_0_1, p1 ); + __m128 p2_d = _mm_mul_ps( d_2_3, p2 ); + __m128 p3_d = _mm_mul_ps( d_2_3, p3 ); - uint32_t qh; - uint64_t tmp[4]; + // Acummulate + acc_0 = _mm_add_ps(p0_d, acc_0); + acc_1 = _mm_add_ps(p1_d, acc_1); + acc_2 = _mm_add_ps(p2_d, acc_2); + acc_3 = _mm_add_ps(p3_d, acc_3); + } - // TODO: check if unrolling this is better - for (int i = 0; i < nb; ++i) { - const block_q5_0 * restrict x0 = &x[i]; - const block_q8_0 * restrict y0 = &y[i]; + *s = hsum_float_4x4(acc_0, acc_1, acc_2, acc_3); +#elif defined(__riscv_v_intrinsic) + float sumf = 0.0; - const v128_t m4b = wasm_i8x16_splat(0x0F); + size_t vl = __riscv_vsetvl_e8m1(qk/2); - // extract the 5th bit - memcpy(&qh, x0->qh, sizeof(qh)); + for (int i = 0; i < nb; i++) { + // load elements + vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[i].qs, vl); - tmp[0] = table_b2b_1[(qh >> 0) & 0xFF]; - tmp[1] = table_b2b_1[(qh >> 8) & 0xFF]; - tmp[2] = table_b2b_1[(qh >> 16) & 0xFF]; - tmp[3] = table_b2b_1[(qh >> 24) ]; + vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[i].qs, vl); + vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[i].qs+16, vl); - const v128_t qhl = wasm_v128_load(tmp + 0); - const v128_t qhh = wasm_v128_load(tmp + 2); + // mask and store lower part of x, and then upper part + vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - const v128_t v0 = wasm_v128_load(x0->qs); + vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); - // 4-bit -> 8-bit - const v128_t v0l = wasm_v128_and (v0, m4b); - const v128_t v0h = wasm_u8x16_shr(v0, 4); + // subtract offset + vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 8, vl); + vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 8, vl); - // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) - const v128_t v0lf = wasm_i8x16_sub(v0l, qhl); - const v128_t v0hf = wasm_i8x16_sub(v0h, qhh); + vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); + vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); - // load y - const v128_t v1l = wasm_v128_load(y0->qs); - const v128_t v1h = wasm_v128_load(y0->qs + 16); + vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); - // int8x16 -> int16x8 - const v128_t v0lfl = wasm_i16x8_extend_low_i8x16 (v0lf); - const v128_t v0lfh = wasm_i16x8_extend_high_i8x16(v0lf); - const v128_t v0hfl = wasm_i16x8_extend_low_i8x16 (v0hf); - const v128_t v0hfh = wasm_i16x8_extend_high_i8x16(v0hf); + vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); + vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); - const v128_t v1ll = wasm_i16x8_extend_low_i8x16 (v1l); - const v128_t v1lh = wasm_i16x8_extend_high_i8x16(v1l); - const v128_t v1hl = wasm_i16x8_extend_low_i8x16 (v1h); - const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); + int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - // dot product - sumv = wasm_f32x4_add(sumv, wasm_f32x4_mul(wasm_f32x4_convert_i32x4( - wasm_i32x4_add( - wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), - wasm_i32x4_dot_i16x8(v0lfh, v1lh)), - wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), - wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), - wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); + sumf += sumi*GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d); } - *s = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + - wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3); -#elif defined(__AVX2__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); + *s = sumf; +#else + // scalar + float sumf = 0.0; - // Main loop for (int i = 0; i < nb; i++) { - /* Compute combined scale for the block */ - const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d)); - - __m256i bx = bytes_from_nibbles_32(x[i].qs); - __m256i bxhi = bytes_from_bits_32(x[i].qh); - bxhi = _mm256_andnot_si256(bxhi, _mm256_set1_epi8((char)0xF0)); - bx = _mm256_or_si256(bx, bxhi); + int sumi = 0; - __m256i by = _mm256_loadu_si256((const __m256i *)y[i].qs); + for (int j = 0; j < qk/2; ++j) { + const int v0 = (x[i].qs[j] & 0x0F) - 8; + const int v1 = (x[i].qs[j] >> 4) - 8; - const __m256 q = mul_sum_i8_pairs_float(bx, by); + sumi += (v0 * y[i].qs[j]) + (v1 * y[i].qs[j + qk/2]); + } - /* Multiply q with scale and accumulate */ - acc = _mm256_fmadd_ps(d, q, acc); + sumf += sumi*GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d); } - *s = hsum_float_8(acc); -#elif defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); - __m128i mask = _mm_set1_epi8((char)0xF0); + *s = sumf; +#endif +} - // Main loop - for (int i = 0; i < nb; i++) { - /* Compute combined scale for the block */ - const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d)); +void ggml_vec_dot_q4_1_q8_1(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_1; + const int nb = n / qk; - __m256i bx = bytes_from_nibbles_32(x[i].qs); - const __m256i bxhi = bytes_from_bits_32(x[i].qh); - __m128i bxhil = _mm256_castsi256_si128(bxhi); - __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); - bxhil = _mm_andnot_si128(bxhil, mask); - bxhih = _mm_andnot_si128(bxhih, mask); - __m128i bxl = _mm256_castsi256_si128(bx); - __m128i bxh = _mm256_extractf128_si256(bx, 1); - bxl = _mm_or_si128(bxl, bxhil); - bxh = _mm_or_si128(bxh, bxhih); - bx = MM256_SET_M128I(bxh, bxl); + assert(n % qk == 0); +#if defined(__ARM_FEATURE_MATMUL_INT8) + assert((nrc == 2) || (nrc == 1)); +#else + assert(nrc == 1); +#endif + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const __m256i by = _mm256_loadu_si256((const __m256i *)y[i].qs); + const block_q4_1 * restrict x = vx; + const block_q8_1 * restrict y = vy; - const __m256 q = mul_sum_i8_pairs_float(bx, by); +#if defined(__ARM_FEATURE_MATMUL_INT8) + if (nrc == 2) { + const block_q4_1 * restrict vx0 = vx; + const block_q4_1 * restrict vx1 = vx + bx; + const block_q8_1 * restrict vy0 = vy; + const block_q8_1 * restrict vy1 = vy + by; - /* Multiply q with scale and accumulate */ - acc = _mm256_add_ps(_mm256_mul_ps(d, q), acc); - } + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t summs0 = vdupq_n_f32(0.0f); - *s = hsum_float_8(acc); -#elif defined(__riscv_v_intrinsic) - float sumf = 0.0; + for (int i = 0; i < nb; i++) { + const block_q4_1 * restrict b_x0 = &vx0[i]; + const block_q4_1 * restrict b_x1 = &vx1[i]; + const block_q8_1 * restrict b_y0 = &vy0[i]; + const block_q8_1 * restrict b_y1 = &vy1[i]; - uint32_t qh; + float32x4_t summs_t = {GGML_FP16_TO_FP32(b_x0->m) * GGML_FP16_TO_FP32(b_y0->s), + GGML_FP16_TO_FP32(b_x1->m) * GGML_FP16_TO_FP32(b_y0->s), + GGML_FP16_TO_FP32(b_x0->m) * GGML_FP16_TO_FP32(b_y1->s), + GGML_FP16_TO_FP32(b_x1->m) * GGML_FP16_TO_FP32(b_y1->s)}; + summs0 += summs_t; - size_t vl = __riscv_vsetvl_e8m1(qk/2); + const uint8x16_t m4b = vdupq_n_u8(0x0F); - // These tempory registers are for masking and shift operations - vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); - vuint32m2_t vt_2 = __riscv_vsll_vv_u32m2(__riscv_vmv_v_x_u32m2(1, vl), vt_1, vl); + const uint8x16_t v0_0 = vld1q_u8(b_x0->qs); + const uint8x16_t v0_1 = vld1q_u8(b_x1->qs); - vuint32m2_t vt_3 = __riscv_vsll_vx_u32m2(vt_2, 16, vl); - vuint32m2_t vt_4 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); + // 4-bit -> 8-bit + const int8x16_t x0_l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t x0_h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t x1_l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t x1_h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - for (int i = 0; i < nb; i++) { - memcpy(&qh, x[i].qh, sizeof(uint32_t)); + // load y + const int8x16_t y0_l = vld1q_s8(b_y0->qs); + const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); + const int8x16_t y1_l = vld1q_s8(b_y1->qs); + const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); - // ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; - vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(vt_2, qh, vl); - vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(xha_0, vt_1, vl); - vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); + // mmla into int32x4_t + float32x4_t scale = {GGML_FP16_TO_FP32(b_x0->d)*b_y0->d, + GGML_FP16_TO_FP32(b_x0->d)*b_y1->d, + GGML_FP16_TO_FP32(b_x1->d)*b_y0->d, + GGML_FP16_TO_FP32(b_x1->d)*b_y1->d}; - // ((qh & (1u << (j + 16))) >> (j + 12)); - vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(vt_3, qh, vl); - vuint32m2_t xhl_1 = __riscv_vsrl_vv_u32m2(xha_1, vt_4, vl); + int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); - // narrowing - vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xhl_0, vl); - vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); + int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xhl_1, vl); - vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); + int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - // load + int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), + l1, r1)), l2, r2)), l3, r3))), scale); + } + + float32x4_t sumv1 = vextq_f32(sumv0, sumv0, 2); + float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); + sumv2 = sumv2 + summs0; + + vst1_f32(s, vget_low_f32(sumv2)); + vst1_f32(s + bs, vget_high_f32(sumv2)); + return; + } +#endif + // TODO: add WASM SIMD +#if defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); + + float summs = 0; + + assert(nb % 2 == 0); // TODO: handle odd nb + + for (int i = 0; i < nb; i += 2) { + const block_q4_1 * restrict x0 = &x[i + 0]; + const block_q4_1 * restrict x1 = &x[i + 1]; + const block_q8_1 * restrict y0 = &y[i + 0]; + const block_q8_1 * restrict y1 = &y[i + 1]; + + summs += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s) + GGML_FP16_TO_FP32(x1->m) * GGML_FP16_TO_FP32(y1->s); + + const uint8x16_t m4b = vdupq_n_u8(0x0F); + + const uint8x16_t v0_0 = vld1q_u8(x0->qs); + const uint8x16_t v0_1 = vld1q_u8(x1->qs); + + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + + // load y + const int8x16_t v1_0l = vld1q_s8(y0->qs); + const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); + const int8x16_t v1_1l = vld1q_s8(y1->qs); + const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); + + // dot product into int32x4_t + const int32x4_t p_0 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_0l, v1_0l), v0_0h, v1_0h); + const int32x4_t p_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), v0_1l, v1_1l), v0_1h, v1_1h); + + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(p_0), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(p_1), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } + + *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs; +#elif defined(__AVX2__) || defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + + float summs = 0; + + // Main loop + for (int i = 0; i < nb; ++i) { + const float d0 = GGML_FP16_TO_FP32(x[i].d); + const float d1 = GGML_FP16_TO_FP32(y[i].d); + + summs += GGML_FP16_TO_FP32(x[i].m) * GGML_FP16_TO_FP32(y[i].s); + + const __m256 d0v = _mm256_set1_ps( d0 ); + const __m256 d1v = _mm256_set1_ps( d1 ); + + // Compute combined scales + const __m256 d0d1 = _mm256_mul_ps( d0v, d1v ); + + // Load 16 bytes, and unpack 4 bit fields into bytes, making 32 bytes + const __m256i qx = bytes_from_nibbles_32(x[i].qs); + const __m256i qy = _mm256_loadu_si256( (const __m256i *)y[i].qs ); + + const __m256 xy = mul_sum_us8_pairs_float(qx, qy); + + // Accumulate d0*d1*x*y +#if defined(__AVX2__) + acc = _mm256_fmadd_ps( d0d1, xy, acc ); +#else + acc = _mm256_add_ps( _mm256_mul_ps( d0d1, xy ), acc ); +#endif + } + + *s = hsum_float_8(acc) + summs; +#elif defined(__riscv_v_intrinsic) + float sumf = 0.0; + + size_t vl = __riscv_vsetvl_e8m1(qk/2); + + for (int i = 0; i < nb; i++) { + // load elements vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[i].qs, vl); vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[i].qs, vl); vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[i].qs+16, vl); - vuint8mf2_t x_at = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); - vuint8mf2_t x_lt = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - - vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); - vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); - - vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + // mask and store lower part of x, and then upper part + vuint8mf2_t x_a = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_l = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 16, vl); - vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 16, vl); + vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); @@ -3068,7 +4241,7 @@ void ggml_vec_dot_q5_0_q8_0(const int n, float * restrict s, const void * restri int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)) * sumi; + sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d))*sumi + GGML_FP16_TO_FP32(x[i].m)*GGML_FP16_TO_FP32(y[i].s); } *s = sumf; @@ -3077,45 +4250,41 @@ void ggml_vec_dot_q5_0_q8_0(const int n, float * restrict s, const void * restri float sumf = 0.0; for (int i = 0; i < nb; i++) { - uint32_t qh; - memcpy(&qh, x[i].qh, sizeof(qh)); - int sumi = 0; for (int j = 0; j < qk/2; ++j) { - const uint8_t xh_0 = ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; - const uint8_t xh_1 = ((qh & (1u << (j + 16))) >> (j + 12)); - - const int32_t x0 = ((x[i].qs[j] & 0x0F) | xh_0) - 16; - const int32_t x1 = ((x[i].qs[j] >> 4) | xh_1) - 16; + const int v0 = (x[i].qs[j] & 0x0F); + const int v1 = (x[i].qs[j] >> 4); - sumi += (x0 * y[i].qs[j]) + (x1 * y[i].qs[j + qk/2]); + sumi += (v0 * y[i].qs[j]) + (v1 * y[i].qs[j + qk/2]); } - sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)) * sumi; + sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d))*sumi + GGML_FP16_TO_FP32(x[i].m)*GGML_FP16_TO_FP32(y[i].s); } *s = sumf; #endif } -void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int qk = QK8_1; +void ggml_vec_dot_q5_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_0; const int nb = n / qk; assert(n % qk == 0); - assert(qk == QK5_1); + assert(qk == QK5_0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q5_1 * restrict x = vx; - const block_q8_1 * restrict y = vy; + const block_q5_0 * restrict x = vx; + const block_q8_0 * restrict y = vy; #if defined(__ARM_NEON) float32x4_t sumv0 = vdupq_n_f32(0.0f); float32x4_t sumv1 = vdupq_n_f32(0.0f); - float summs0 = 0.0f; - float summs1 = 0.0f; - uint32_t qh0; uint32_t qh1; @@ -3125,29 +4294,26 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri assert(nb % 2 == 0); // TODO: handle odd nb for (int i = 0; i < nb; i += 2) { - const block_q5_1 * restrict x0 = &x[i]; - const block_q5_1 * restrict x1 = &x[i + 1]; - const block_q8_1 * restrict y0 = &y[i]; - const block_q8_1 * restrict y1 = &y[i + 1]; + const block_q5_0 * restrict x0 = &x[i]; + const block_q5_0 * restrict x1 = &x[i + 1]; + const block_q8_0 * restrict y0 = &y[i]; + const block_q8_0 * restrict y1 = &y[i + 1]; const uint8x16_t m4b = vdupq_n_u8(0x0F); - summs0 += GGML_FP16_TO_FP32(x0->m) * y0->s; - summs1 += GGML_FP16_TO_FP32(x1->m) * y1->s; - - // extract the 5th bit via lookup table ((b) << 4) + // extract the 5th bit via lookup table ((!b) << 4) memcpy(&qh0, x0->qh, sizeof(qh0)); memcpy(&qh1, x1->qh, sizeof(qh1)); - tmp0[0] = table_b2b_0[(qh0 >> 0) & 0xFF]; - tmp0[1] = table_b2b_0[(qh0 >> 8) & 0xFF]; - tmp0[2] = table_b2b_0[(qh0 >> 16) & 0xFF]; - tmp0[3] = table_b2b_0[(qh0 >> 24) ]; + tmp0[0] = table_b2b_1[(qh0 >> 0) & 0xFF]; + tmp0[1] = table_b2b_1[(qh0 >> 8) & 0xFF]; + tmp0[2] = table_b2b_1[(qh0 >> 16) & 0xFF]; + tmp0[3] = table_b2b_1[(qh0 >> 24) ]; - tmp1[0] = table_b2b_0[(qh1 >> 0) & 0xFF]; - tmp1[1] = table_b2b_0[(qh1 >> 8) & 0xFF]; - tmp1[2] = table_b2b_0[(qh1 >> 16) & 0xFF]; - tmp1[3] = table_b2b_0[(qh1 >> 24) ]; + tmp1[0] = table_b2b_1[(qh1 >> 0) & 0xFF]; + tmp1[1] = table_b2b_1[(qh1 >> 8) & 0xFF]; + tmp1[2] = table_b2b_1[(qh1 >> 16) & 0xFF]; + tmp1[3] = table_b2b_1[(qh1 >> 24) ]; const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); @@ -3158,16 +4324,16 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri const uint8x16_t v0_1 = vld1q_u8(x1->qs); // 4-bit -> 8-bit - const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); - const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); - const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); - const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); + int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - // add high bit - const int8x16_t v0_0lf = vorrq_s8(v0_0l, qhl0); - const int8x16_t v0_0hf = vorrq_s8(v0_0h, qhh0); - const int8x16_t v0_1lf = vorrq_s8(v0_1l, qhl1); - const int8x16_t v0_1hf = vorrq_s8(v0_1h, qhh1); + // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) + const int8x16_t v0_0lf = vsubq_s8(v0_0l, qhl0); + const int8x16_t v0_0hf = vsubq_s8(v0_0h, qhh0); + const int8x16_t v0_1lf = vsubq_s8(v0_1l, qhl1); + const int8x16_t v0_1hf = vsubq_s8(v0_1h, qhh1); // load y const int8x16_t v1_0l = vld1q_s8(y0->qs); @@ -3175,59 +4341,35 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri const int8x16_t v1_1l = vld1q_s8(y1->qs); const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); -#if defined(__ARM_FEATURE_DOTPROD) sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( - vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), - vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*y0->d); + ggml_vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( - vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), - vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*y1->d); -#else - const int16x8_t pl0l = vmull_s8(vget_low_s8 (v0_0lf), vget_low_s8 (v1_0l)); - const int16x8_t pl0h = vmull_s8(vget_high_s8(v0_0lf), vget_high_s8(v1_0l)); - const int16x8_t ph0l = vmull_s8(vget_low_s8 (v0_0hf), vget_low_s8 (v1_0h)); - const int16x8_t ph0h = vmull_s8(vget_high_s8(v0_0hf), vget_high_s8(v1_0h)); - - const int16x8_t pl1l = vmull_s8(vget_low_s8 (v0_1lf), vget_low_s8 (v1_1l)); - const int16x8_t pl1h = vmull_s8(vget_high_s8(v0_1lf), vget_high_s8(v1_1l)); - const int16x8_t ph1l = vmull_s8(vget_low_s8 (v0_1hf), vget_low_s8 (v1_1h)); - const int16x8_t ph1h = vmull_s8(vget_high_s8(v0_1hf), vget_high_s8(v1_1h)); - - const int32x4_t pl0 = vaddq_s32(vpaddlq_s16(pl0l), vpaddlq_s16(pl0h)); - const int32x4_t ph0 = vaddq_s32(vpaddlq_s16(ph0l), vpaddlq_s16(ph0h)); - const int32x4_t pl1 = vaddq_s32(vpaddlq_s16(pl1l), vpaddlq_s16(pl1h)); - const int32x4_t ph1 = vaddq_s32(vpaddlq_s16(ph1l), vpaddlq_s16(ph1h)); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32(pl0, ph0)), GGML_FP16_TO_FP32(x0->d)*y0->d); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32(pl1, ph1)), GGML_FP16_TO_FP32(x1->d)*y1->d); -#endif + ggml_vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); } - *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs0 + summs1; + *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); #elif defined(__wasm_simd128__) v128_t sumv = wasm_f32x4_splat(0.0f); - float summs = 0.0f; - uint32_t qh; uint64_t tmp[4]; // TODO: check if unrolling this is better for (int i = 0; i < nb; ++i) { - const block_q5_1 * restrict x0 = &x[i]; - const block_q8_1 * restrict y0 = &y[i]; - - summs += GGML_FP16_TO_FP32(x0->m) * y0->s; + const block_q5_0 * restrict x0 = &x[i]; + const block_q8_0 * restrict y0 = &y[i]; - const v128_t m4b = wasm_i8x16_splat(0x0F); + const v128_t m4b = wasm_i8x16_splat(0x0F); // extract the 5th bit memcpy(&qh, x0->qh, sizeof(qh)); - tmp[0] = table_b2b_0[(qh >> 0) & 0xFF]; - tmp[1] = table_b2b_0[(qh >> 8) & 0xFF]; - tmp[2] = table_b2b_0[(qh >> 16) & 0xFF]; - tmp[3] = table_b2b_0[(qh >> 24) ]; + tmp[0] = table_b2b_1[(qh >> 0) & 0xFF]; + tmp[1] = table_b2b_1[(qh >> 8) & 0xFF]; + tmp[2] = table_b2b_1[(qh >> 16) & 0xFF]; + tmp[3] = table_b2b_1[(qh >> 24) ]; const v128_t qhl = wasm_v128_load(tmp + 0); const v128_t qhh = wasm_v128_load(tmp + 2); @@ -3238,9 +4380,9 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri const v128_t v0l = wasm_v128_and (v0, m4b); const v128_t v0h = wasm_u8x16_shr(v0, 4); - // add high bit - const v128_t v0lf = wasm_v128_or(v0l, qhl); - const v128_t v0hf = wasm_v128_or(v0h, qhh); + // add high bit and sub 16 (equivalent to sub 0x10 when bit is zero) + const v128_t v0lf = wasm_i8x16_sub(v0l, qhl); + const v128_t v0hf = wasm_i8x16_sub(v0h, qhh); // load y const v128_t v1l = wasm_v128_load(y0->qs); @@ -3258,77 +4400,71 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); // dot product - sumv = wasm_f32x4_add(sumv, - wasm_f32x4_mul(wasm_f32x4_convert_i32x4(wasm_i32x4_add( + sumv = wasm_f32x4_add(sumv, wasm_f32x4_mul(wasm_f32x4_convert_i32x4( + wasm_i32x4_add( wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), wasm_i32x4_dot_i16x8(v0lfh, v1lh)), wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), - wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * y0->d))); + wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); } *s = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + - wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3) + summs; + wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3); #elif defined(__AVX2__) // Initialize accumulator with zeros __m256 acc = _mm256_setzero_ps(); - float summs = 0.0f; - // Main loop for (int i = 0; i < nb; i++) { - const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d)); - - summs += GGML_FP16_TO_FP32(x[i].m) * y[i].s; + /* Compute combined scale for the block */ + const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d)); - __m256i bx = bytes_from_nibbles_32(x[i].qs); + __m256i qx = bytes_from_nibbles_32(x[i].qs); __m256i bxhi = bytes_from_bits_32(x[i].qh); - bxhi = _mm256_and_si256(bxhi, _mm256_set1_epi8(0x10)); - bx = _mm256_or_si256(bx, bxhi); + bxhi = _mm256_andnot_si256(bxhi, _mm256_set1_epi8((char)0xF0)); + qx = _mm256_or_si256(qx, bxhi); - const __m256 dy = _mm256_set1_ps(y[i].d); - const __m256i by = _mm256_loadu_si256((const __m256i *)y[i].qs); + __m256i qy = _mm256_loadu_si256((const __m256i *)y[i].qs); - const __m256 q = mul_sum_us8_pairs_float(bx, by); + const __m256 q = mul_sum_i8_pairs_float(qx, qy); - acc = _mm256_fmadd_ps(q, _mm256_mul_ps(dx, dy), acc); + /* Multiply q with scale and accumulate */ + acc = _mm256_fmadd_ps(d, q, acc); } - *s = hsum_float_8(acc) + summs; + *s = hsum_float_8(acc); #elif defined(__AVX__) // Initialize accumulator with zeros __m256 acc = _mm256_setzero_ps(); - __m128i mask = _mm_set1_epi8(0x10); - - float summs = 0.0f; + __m128i mask = _mm_set1_epi8((char)0xF0); // Main loop for (int i = 0; i < nb; i++) { - const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d)); - - summs += GGML_FP16_TO_FP32(x[i].m) * y[i].s; + /* Compute combined scale for the block */ + const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d)); - __m256i bx = bytes_from_nibbles_32(x[i].qs); + __m256i bx_0 = bytes_from_nibbles_32(x[i].qs); const __m256i bxhi = bytes_from_bits_32(x[i].qh); __m128i bxhil = _mm256_castsi256_si128(bxhi); __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); - bxhil = _mm_and_si128(bxhil, mask); - bxhih = _mm_and_si128(bxhih, mask); - __m128i bxl = _mm256_castsi256_si128(bx); - __m128i bxh = _mm256_extractf128_si256(bx, 1); + bxhil = _mm_andnot_si128(bxhil, mask); + bxhih = _mm_andnot_si128(bxhih, mask); + __m128i bxl = _mm256_castsi256_si128(bx_0); + __m128i bxh = _mm256_extractf128_si256(bx_0, 1); bxl = _mm_or_si128(bxl, bxhil); bxh = _mm_or_si128(bxh, bxhih); - bx = MM256_SET_M128I(bxh, bxl); + bx_0 = MM256_SET_M128I(bxh, bxl); - const __m256 dy = _mm256_set1_ps(y[i].d); - const __m256i by = _mm256_loadu_si256((const __m256i *)y[i].qs); + const __m256i by_0 = _mm256_loadu_si256((const __m256i *)y[i].qs); - const __m256 q = mul_sum_us8_pairs_float(bx, by); + const __m256 q = mul_sum_i8_pairs_float(bx_0, by_0); - acc = _mm256_add_ps(_mm256_mul_ps(q, _mm256_mul_ps(dx, dy)), acc); + /* Multiply q with scale and accumulate */ + acc = _mm256_add_ps(_mm256_mul_ps(d, q), acc); } - *s = hsum_float_8(acc) + summs; + *s = hsum_float_8(acc); #elif defined(__riscv_v_intrinsic) float sumf = 0.0; @@ -3336,30 +4472,30 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri size_t vl = __riscv_vsetvl_e8m1(qk/2); - // temporary registers for shift operations + // These temporary registers are for masking and shift operations vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); - vuint32m2_t vt_2 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); + vuint32m2_t vt_2 = __riscv_vsll_vv_u32m2(__riscv_vmv_v_x_u32m2(1, vl), vt_1, vl); + + vuint32m2_t vt_3 = __riscv_vsll_vx_u32m2(vt_2, 16, vl); + vuint32m2_t vt_4 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); for (int i = 0; i < nb; i++) { memcpy(&qh, x[i].qh, sizeof(uint32_t)); - // load qh - vuint32m2_t vqh = __riscv_vmv_v_x_u32m2(qh, vl); + // ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; + vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(vt_2, qh, vl); + vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(xha_0, vt_1, vl); + vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); - // ((qh >> (j + 0)) << 4) & 0x10; - vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(vqh, vt_1, vl); - vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); - vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(xhl_0, 0x10, vl); - - // ((qh >> (j + 12)) ) & 0x10; - vuint32m2_t xhr_1 = __riscv_vsrl_vv_u32m2(vqh, vt_2, vl); - vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(xhr_1, 0x10, vl); + // ((qh & (1u << (j + 16))) >> (j + 12)); + vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(vt_3, qh, vl); + vuint32m2_t xhl_1 = __riscv_vsrl_vv_u32m2(xha_1, vt_4, vl); // narrowing - vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xha_0, vl); + vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xhl_0, vl); vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); - vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xha_1, vl); + vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xhl_1, vl); vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); // load @@ -3374,8 +4510,11 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); - vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); - vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + vint8mf2_t x_ai = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t x_li = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); + + vint8mf2_t v0 = __riscv_vsub_vx_i8mf2(x_ai, 16, vl); + vint8mf2_t v1 = __riscv_vsub_vx_i8mf2(x_li, 16, vl); vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); @@ -3387,7 +4526,7 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); - sumf += (GGML_FP16_TO_FP32(x[i].d)*y[i].d)*sumi + GGML_FP16_TO_FP32(x[i].m)*y[i].s; + sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)) * sumi; } *s = sumf; @@ -3402,540 +4541,499 @@ void ggml_vec_dot_q5_1_q8_1(const int n, float * restrict s, const void * restri int sumi = 0; for (int j = 0; j < qk/2; ++j) { - const uint8_t xh_0 = ((qh >> (j + 0)) << 4) & 0x10; - const uint8_t xh_1 = ((qh >> (j + 12)) ) & 0x10; + const uint8_t xh_0 = ((qh & (1u << (j + 0 ))) >> (j + 0 )) << 4; + const uint8_t xh_1 = ((qh & (1u << (j + 16))) >> (j + 12)); - const int32_t x0 = (x[i].qs[j] & 0xF) | xh_0; - const int32_t x1 = (x[i].qs[j] >> 4) | xh_1; + const int32_t x0 = ((x[i].qs[j] & 0x0F) | xh_0) - 16; + const int32_t x1 = ((x[i].qs[j] >> 4) | xh_1) - 16; sumi += (x0 * y[i].qs[j]) + (x1 * y[i].qs[j + qk/2]); } - sumf += (GGML_FP16_TO_FP32(x[i].d)*y[i].d)*sumi + GGML_FP16_TO_FP32(x[i].m)*y[i].s; + sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)) * sumi; } *s = sumf; #endif } -void ggml_vec_dot_q8_0_q8_0(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - const int qk = QK8_0; +void ggml_vec_dot_q5_1_q8_1(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_1; const int nb = n / qk; assert(n % qk == 0); + assert(qk == QK5_1); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q8_0 * restrict x = vx; - const block_q8_0 * restrict y = vy; + const block_q5_1 * restrict x = vx; + const block_q8_1 * restrict y = vy; #if defined(__ARM_NEON) float32x4_t sumv0 = vdupq_n_f32(0.0f); float32x4_t sumv1 = vdupq_n_f32(0.0f); - assert(nb % 2 == 0); // TODO: handle odd nb - - for (int i = 0; i < nb; i += 2) { - const block_q8_0 * restrict x0 = &x[i + 0]; - const block_q8_0 * restrict x1 = &x[i + 1]; - const block_q8_0 * restrict y0 = &y[i + 0]; - const block_q8_0 * restrict y1 = &y[i + 1]; + float summs0 = 0.0f; + float summs1 = 0.0f; - const int8x16_t x0_0 = vld1q_s8(x0->qs); - const int8x16_t x0_1 = vld1q_s8(x0->qs + 16); - const int8x16_t x1_0 = vld1q_s8(x1->qs); - const int8x16_t x1_1 = vld1q_s8(x1->qs + 16); + uint32_t qh0; + uint32_t qh1; - // load y - const int8x16_t y0_0 = vld1q_s8(y0->qs); - const int8x16_t y0_1 = vld1q_s8(y0->qs + 16); - const int8x16_t y1_0 = vld1q_s8(y1->qs); - const int8x16_t y1_1 = vld1q_s8(y1->qs + 16); + uint64_t tmp0[4]; + uint64_t tmp1[4]; -#if defined(__ARM_FEATURE_DOTPROD) - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( - vdotq_s32(vdupq_n_s32(0), x0_0, y0_0), - vdotq_s32(vdupq_n_s32(0), x0_1, y0_1))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + assert(nb % 2 == 0); // TODO: handle odd nb - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( - vdotq_s32(vdupq_n_s32(0), x1_0, y1_0), - vdotq_s32(vdupq_n_s32(0), x1_1, y1_1))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + for (int i = 0; i < nb; i += 2) { + const block_q5_1 * restrict x0 = &x[i]; + const block_q5_1 * restrict x1 = &x[i + 1]; + const block_q8_1 * restrict y0 = &y[i]; + const block_q8_1 * restrict y1 = &y[i + 1]; -#else - const int16x8_t p0_0 = vmull_s8(vget_low_s8 (x0_0), vget_low_s8 (y0_0)); - const int16x8_t p0_1 = vmull_s8(vget_high_s8(x0_0), vget_high_s8(y0_0)); - const int16x8_t p0_2 = vmull_s8(vget_low_s8 (x0_1), vget_low_s8 (y0_1)); - const int16x8_t p0_3 = vmull_s8(vget_high_s8(x0_1), vget_high_s8(y0_1)); - - const int16x8_t p1_0 = vmull_s8(vget_low_s8 (x1_0), vget_low_s8 (y1_0)); - const int16x8_t p1_1 = vmull_s8(vget_high_s8(x1_0), vget_high_s8(y1_0)); - const int16x8_t p1_2 = vmull_s8(vget_low_s8 (x1_1), vget_low_s8 (y1_1)); - const int16x8_t p1_3 = vmull_s8(vget_high_s8(x1_1), vget_high_s8(y1_1)); - - const int32x4_t p0 = vaddq_s32(vpaddlq_s16(p0_0), vpaddlq_s16(p0_1)); - const int32x4_t p1 = vaddq_s32(vpaddlq_s16(p0_2), vpaddlq_s16(p0_3)); - const int32x4_t p2 = vaddq_s32(vpaddlq_s16(p1_0), vpaddlq_s16(p1_1)); - const int32x4_t p3 = vaddq_s32(vpaddlq_s16(p1_2), vpaddlq_s16(p1_3)); - - sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32(p0, p1)), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32(p2, p3)), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); -#endif - } + const uint8x16_t m4b = vdupq_n_u8(0x0F); - *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); -#elif defined(__AVX2__) || defined(__AVX__) - // Initialize accumulator with zeros - __m256 acc = _mm256_setzero_ps(); + summs0 += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s); + summs1 += GGML_FP16_TO_FP32(x1->m) * GGML_FP16_TO_FP32(y1->s); - // Main loop - for (int i = 0; i < nb; ++i) { - // Compute combined scale for the block - const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d)); - __m256i bx = _mm256_loadu_si256((const __m256i *)x[i].qs); - __m256i by = _mm256_loadu_si256((const __m256i *)y[i].qs); + // extract the 5th bit via lookup table ((b) << 4) + memcpy(&qh0, x0->qh, sizeof(qh0)); + memcpy(&qh1, x1->qh, sizeof(qh1)); - const __m256 q = mul_sum_i8_pairs_float(bx, by); + tmp0[0] = table_b2b_0[(qh0 >> 0) & 0xFF]; + tmp0[1] = table_b2b_0[(qh0 >> 8) & 0xFF]; + tmp0[2] = table_b2b_0[(qh0 >> 16) & 0xFF]; + tmp0[3] = table_b2b_0[(qh0 >> 24) ]; - // Multiply q with scale and accumulate -#if defined(__AVX2__) - acc = _mm256_fmadd_ps( d, q, acc ); -#else - acc = _mm256_add_ps( _mm256_mul_ps( d, q ), acc ); -#endif - } + tmp1[0] = table_b2b_0[(qh1 >> 0) & 0xFF]; + tmp1[1] = table_b2b_0[(qh1 >> 8) & 0xFF]; + tmp1[2] = table_b2b_0[(qh1 >> 16) & 0xFF]; + tmp1[3] = table_b2b_0[(qh1 >> 24) ]; - *s = hsum_float_8(acc); -#elif defined(__riscv_v_intrinsic) - float sumf = 0.0; - size_t vl = __riscv_vsetvl_e8m1(qk); + const int8x16_t qhl0 = vld1q_s8((const int8_t *)(tmp0 + 0)); + const int8x16_t qhh0 = vld1q_s8((const int8_t *)(tmp0 + 2)); + const int8x16_t qhl1 = vld1q_s8((const int8_t *)(tmp1 + 0)); + const int8x16_t qhh1 = vld1q_s8((const int8_t *)(tmp1 + 2)); - for (int i = 0; i < nb; i++) { - // load elements - vint8m1_t bx = __riscv_vle8_v_i8m1(x[i].qs, vl); - vint8m1_t by = __riscv_vle8_v_i8m1(y[i].qs, vl); + const uint8x16_t v0_0 = vld1q_u8(x0->qs); + const uint8x16_t v0_1 = vld1q_u8(x1->qs); - vint16m2_t vw_mul = __riscv_vwmul_vv_i16m2(bx, by, vl); + // 4-bit -> 8-bit + const int8x16_t v0_0l = vreinterpretq_s8_u8(vandq_u8 (v0_0, m4b)); + const int8x16_t v0_0h = vreinterpretq_s8_u8(vshrq_n_u8(v0_0, 4)); + const int8x16_t v0_1l = vreinterpretq_s8_u8(vandq_u8 (v0_1, m4b)); + const int8x16_t v0_1h = vreinterpretq_s8_u8(vshrq_n_u8(v0_1, 4)); - vint32m1_t v_zero = __riscv_vmv_v_x_i32m1(0, vl); - vint32m1_t v_sum = __riscv_vwredsum_vs_i16m2_i32m1(vw_mul, v_zero, vl); + // add high bit + const int8x16_t v0_0lf = vorrq_s8(v0_0l, qhl0); + const int8x16_t v0_0hf = vorrq_s8(v0_0h, qhh0); + const int8x16_t v0_1lf = vorrq_s8(v0_1l, qhl1); + const int8x16_t v0_1hf = vorrq_s8(v0_1h, qhh1); - int sumi = __riscv_vmv_x_s_i32m1_i32(v_sum); + // load y + const int8x16_t v1_0l = vld1q_s8(y0->qs); + const int8x16_t v1_0h = vld1q_s8(y0->qs + 16); + const int8x16_t v1_1l = vld1q_s8(y1->qs); + const int8x16_t v1_1h = vld1q_s8(y1->qs + 16); - sumf += sumi*(GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)); + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), v0_0lf, v1_0l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_0hf, v1_0h))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), v0_1lf, v1_1l), + ggml_vdotq_s32(vdupq_n_s32(0), v0_1hf, v1_1h))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); } - *s = sumf; -#else - // scalar - float sumf = 0.0; + *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1) + summs0 + summs1; +#elif defined(__wasm_simd128__) + v128_t sumv = wasm_f32x4_splat(0.0f); - for (int i = 0; i < nb; i++) { - int sumi = 0; + float summs = 0.0f; - for (int j = 0; j < qk; j++) { - sumi += x[i].qs[j]*y[i].qs[j]; - } + uint32_t qh; + uint64_t tmp[4]; - sumf += sumi*(GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)); - } + // TODO: check if unrolling this is better + for (int i = 0; i < nb; ++i) { + const block_q5_1 * restrict x0 = &x[i]; + const block_q8_1 * restrict y0 = &y[i]; - *s = sumf; -#endif -} + summs += GGML_FP16_TO_FP32(x0->m) * GGML_FP16_TO_FP32(y0->s); -#if QK_K == 256 -void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { + const v128_t m4b = wasm_i8x16_splat(0x0F); - const block_q2_K * restrict x = vx; - const block_q8_K * restrict y = vy; + // extract the 5th bit + memcpy(&qh, x0->qh, sizeof(qh)); - const int nb = n / QK_K; + tmp[0] = table_b2b_0[(qh >> 0) & 0xFF]; + tmp[1] = table_b2b_0[(qh >> 8) & 0xFF]; + tmp[2] = table_b2b_0[(qh >> 16) & 0xFF]; + tmp[3] = table_b2b_0[(qh >> 24) ]; -#ifdef __ARM_NEON + const v128_t qhl = wasm_v128_load(tmp + 0); + const v128_t qhh = wasm_v128_load(tmp + 2); - const uint8x16_t m3 = vdupq_n_u8(0x3); - const uint8x16_t m4 = vdupq_n_u8(0xF); -#if defined(__ARM_FEATURE_DOTPROD) - const int32x4_t vzero = vdupq_n_s32(0); -#endif + const v128_t v0 = wasm_v128_load(x0->qs); - int8x16x2_t q2bytes; - uint8_t aux[16]; + // 4-bit -> 8-bit + const v128_t v0l = wasm_v128_and (v0, m4b); + const v128_t v0h = wasm_u8x16_shr(v0, 4); - float sum = 0; + // add high bit + const v128_t v0lf = wasm_v128_or(v0l, qhl); + const v128_t v0hf = wasm_v128_or(v0h, qhh); - for (int i = 0; i < nb; ++i) { + // load y + const v128_t v1l = wasm_v128_load(y0->qs); + const v128_t v1h = wasm_v128_load(y0->qs + 16); - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + // int8x16 -> int16x8 + const v128_t v0lfl = wasm_i16x8_extend_low_i8x16 (v0lf); + const v128_t v0lfh = wasm_i16x8_extend_high_i8x16(v0lf); + const v128_t v0hfl = wasm_i16x8_extend_low_i8x16 (v0hf); + const v128_t v0hfh = wasm_i16x8_extend_high_i8x16(v0hf); - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - const uint8_t * restrict sc = x[i].scales; + const v128_t v1ll = wasm_i16x8_extend_low_i8x16 (v1l); + const v128_t v1lh = wasm_i16x8_extend_high_i8x16(v1l); + const v128_t v1hl = wasm_i16x8_extend_low_i8x16 (v1h); + const v128_t v1hh = wasm_i16x8_extend_high_i8x16(v1h); - const uint8x16_t mins_and_scales = vld1q_u8(sc); - const uint8x16_t scales = vandq_u8(mins_and_scales, m4); - vst1q_u8(aux, scales); + // dot product + sumv = wasm_f32x4_add(sumv, + wasm_f32x4_mul(wasm_f32x4_convert_i32x4(wasm_i32x4_add( + wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0lfl, v1ll), + wasm_i32x4_dot_i16x8(v0lfh, v1lh)), + wasm_i32x4_add(wasm_i32x4_dot_i16x8(v0hfl, v1hl), + wasm_i32x4_dot_i16x8(v0hfh, v1hh)))), + wasm_f32x4_splat(GGML_FP16_TO_FP32(x0->d) * GGML_FP16_TO_FP32(y0->d)))); + } - const uint8x16_t mins = vshrq_n_u8(mins_and_scales, 4); - const int16x8x2_t q8sums = vld1q_s16_x2(y[i].bsums); - const int16x8x2_t mins16 = {vreinterpretq_s16_u16(vmovl_u8(vget_low_u8(mins))), vreinterpretq_s16_u16(vmovl_u8(vget_high_u8(mins)))}; - const int32x4_t s0 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[0]), vget_low_s16 (q8sums.val[0])), - vmull_s16(vget_high_s16(mins16.val[0]), vget_high_s16(q8sums.val[0]))); - const int32x4_t s1 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[1]), vget_low_s16 (q8sums.val[1])), - vmull_s16(vget_high_s16(mins16.val[1]), vget_high_s16(q8sums.val[1]))); - sum += dmin * vaddvq_s32(vaddq_s32(s0, s1)); + *s = wasm_f32x4_extract_lane(sumv, 0) + wasm_f32x4_extract_lane(sumv, 1) + + wasm_f32x4_extract_lane(sumv, 2) + wasm_f32x4_extract_lane(sumv, 3) + summs; +#elif defined(__AVX2__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); - int isum = 0; - int is = 0; + float summs = 0.0f; -// We use this macro instead of a function call because for some reason -// the code runs 2-3% slower, even if the function is declared inline -#if defined(__ARM_FEATURE_DOTPROD) -#define MULTIPLY_ACCUM_WITH_SCALE(index)\ - isum += vaddvq_s32(vdotq_s32(vzero, q2bytes.val[0], q8bytes.val[0])) * aux[is+(index)];\ - isum += vaddvq_s32(vdotq_s32(vzero, q2bytes.val[1], q8bytes.val[1])) * aux[is+1+(index)]; -#else -#define MULTIPLY_ACCUM_WITH_SCALE(index)\ - {\ - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q2bytes.val[0]), vget_low_s8 (q8bytes.val[0])),\ - vmull_s8(vget_high_s8(q2bytes.val[0]), vget_high_s8(q8bytes.val[0])));\ - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q2bytes.val[1]), vget_low_s8 (q8bytes.val[1])),\ - vmull_s8(vget_high_s8(q2bytes.val[1]), vget_high_s8(q8bytes.val[1])));\ - isum += vaddvq_s16(p1) * aux[is+(index)] + vaddvq_s16(p2) * aux[is+1+(index)];\ - } -#endif + // Main loop + for (int i = 0; i < nb; i++) { + const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d)); -#define SHIFT_MULTIPLY_ACCUM_WITH_SCALE(shift, index)\ - q8bytes = vld1q_s8_x2(q8); q8 += 32;\ - q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[0], (shift)), m3));\ - q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[1], (shift)), m3));\ - MULTIPLY_ACCUM_WITH_SCALE((index)); + summs += GGML_FP16_TO_FP32(x[i].m) * GGML_FP16_TO_FP32(y[i].s); + __m256i qx = bytes_from_nibbles_32(x[i].qs); + __m256i bxhi = bytes_from_bits_32(x[i].qh); + bxhi = _mm256_and_si256(bxhi, _mm256_set1_epi8(0x10)); + qx = _mm256_or_si256(qx, bxhi); - for (int j = 0; j < QK_K/128; ++j) { + const __m256 dy = _mm256_set1_ps(GGML_FP16_TO_FP32(y[i].d)); + const __m256i qy = _mm256_loadu_si256((const __m256i *)y[i].qs); - const uint8x16x2_t q2bits = vld1q_u8_x2(q2); q2 += 32; + const __m256 q = mul_sum_us8_pairs_float(qx, qy); - int8x16x2_t q8bytes = vld1q_s8_x2(q8); q8 += 32; - q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[0], m3)); - q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[1], m3)); - MULTIPLY_ACCUM_WITH_SCALE(0); + acc = _mm256_fmadd_ps(q, _mm256_mul_ps(dx, dy), acc); + } - SHIFT_MULTIPLY_ACCUM_WITH_SCALE(2, 2); + *s = hsum_float_8(acc) + summs; +#elif defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); + __m128i mask = _mm_set1_epi8(0x10); - SHIFT_MULTIPLY_ACCUM_WITH_SCALE(4, 4); + float summs = 0.0f; - SHIFT_MULTIPLY_ACCUM_WITH_SCALE(6, 6); + // Main loop + for (int i = 0; i < nb; i++) { + const __m256 dx = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d)); - is += 8; - } - sum += d * isum; + summs += GGML_FP16_TO_FP32(x[i].m) * GGML_FP16_TO_FP32(y[i].s); - } + __m256i bx_0 = bytes_from_nibbles_32(x[i].qs); + const __m256i bxhi = bytes_from_bits_32(x[i].qh); + __m128i bxhil = _mm256_castsi256_si128(bxhi); + __m128i bxhih = _mm256_extractf128_si256(bxhi, 1); + bxhil = _mm_and_si128(bxhil, mask); + bxhih = _mm_and_si128(bxhih, mask); + __m128i bxl = _mm256_castsi256_si128(bx_0); + __m128i bxh = _mm256_extractf128_si256(bx_0, 1); + bxl = _mm_or_si128(bxl, bxhil); + bxh = _mm_or_si128(bxh, bxhih); + bx_0 = MM256_SET_M128I(bxh, bxl); - *s = sum; + const __m256 dy = _mm256_set1_ps(GGML_FP16_TO_FP32(y[i].d)); + const __m256i by_0 = _mm256_loadu_si256((const __m256i *)y[i].qs); -#elif defined __AVX2__ + const __m256 q = mul_sum_us8_pairs_float(bx_0, by_0); - const __m256i m3 = _mm256_set1_epi8(3); - const __m128i m4 = _mm_set1_epi8(0xF); + acc = _mm256_add_ps(_mm256_mul_ps(q, _mm256_mul_ps(dx, dy)), acc); + } - __m256 acc = _mm256_setzero_ps(); + *s = hsum_float_8(acc) + summs; +#elif defined(__riscv_v_intrinsic) + float sumf = 0.0; - for (int i = 0; i < nb; ++i) { + uint32_t qh; - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + size_t vl = __riscv_vsetvl_e8m1(qk/2); - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; + // temporary registers for shift operations + vuint32m2_t vt_1 = __riscv_vid_v_u32m2(vl); + vuint32m2_t vt_2 = __riscv_vadd_vx_u32m2(vt_1, 12, vl); - const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); - const __m128i scales8 = _mm_and_si128(mins_and_scales, m4); - const __m128i mins8 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); - const __m256i mins = _mm256_cvtepi8_epi16(mins8); - const __m256i prod = _mm256_madd_epi16(mins, _mm256_loadu_si256((const __m256i*)y[i].bsums)); + for (int i = 0; i < nb; i++) { + memcpy(&qh, x[i].qh, sizeof(uint32_t)); - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(prod), acc); + // load qh + vuint32m2_t vqh = __riscv_vmv_v_x_u32m2(qh, vl); - const __m256i all_scales = _mm256_cvtepi8_epi16(scales8); - const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); - const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); - const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; + // ((qh >> (j + 0)) << 4) & 0x10; + vuint32m2_t xhr_0 = __riscv_vsrl_vv_u32m2(vqh, vt_1, vl); + vuint32m2_t xhl_0 = __riscv_vsll_vx_u32m2(xhr_0, 4, vl); + vuint32m2_t xha_0 = __riscv_vand_vx_u32m2(xhl_0, 0x10, vl); - __m256i sumi = _mm256_setzero_si256(); + // ((qh >> (j + 12)) ) & 0x10; + vuint32m2_t xhr_1 = __riscv_vsrl_vv_u32m2(vqh, vt_2, vl); + vuint32m2_t xha_1 = __riscv_vand_vx_u32m2(xhr_1, 0x10, vl); - for (int j = 0; j < QK_K/128; ++j) { + // narrowing + vuint16m1_t xhc_0 = __riscv_vncvt_x_x_w_u16m1(xha_0, vl); + vuint8mf2_t xh_0 = __riscv_vncvt_x_x_w_u8mf2(xhc_0, vl); - const __m256i q2bits = _mm256_loadu_si256((const __m256i*)q2); q2 += 32; + vuint16m1_t xhc_1 = __riscv_vncvt_x_x_w_u16m1(xha_1, vl); + vuint8mf2_t xh_1 = __riscv_vncvt_x_x_w_u8mf2(xhc_1, vl); - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + // load + vuint8mf2_t tx = __riscv_vle8_v_u8mf2(x[i].qs, vl); - const __m256i q2_0 = _mm256_and_si256(q2bits, m3); - const __m256i q2_1 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 2), m3); - const __m256i q2_2 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 4), m3); - const __m256i q2_3 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 6), m3); + vint8mf2_t y0 = __riscv_vle8_v_i8mf2(y[i].qs, vl); + vint8mf2_t y1 = __riscv_vle8_v_i8mf2(y[i].qs+16, vl); - __m256i p0 = _mm256_maddubs_epi16(q2_0, q8_0); - __m256i p1 = _mm256_maddubs_epi16(q2_1, q8_1); - __m256i p2 = _mm256_maddubs_epi16(q2_2, q8_2); - __m256i p3 = _mm256_maddubs_epi16(q2_3, q8_3); + vuint8mf2_t x_at = __riscv_vand_vx_u8mf2(tx, 0x0F, vl); + vuint8mf2_t x_lt = __riscv_vsrl_vx_u8mf2(tx, 0x04, vl); - p0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(0)), p0); - p1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(1)), p1); - p2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(2)), p2); - p3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(3)), p3); + vuint8mf2_t x_a = __riscv_vor_vv_u8mf2(x_at, xh_0, vl); + vuint8mf2_t x_l = __riscv_vor_vv_u8mf2(x_lt, xh_1, vl); - p0 = _mm256_add_epi32(p0, p1); - p2 = _mm256_add_epi32(p2, p3); + vint8mf2_t v0 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_a); + vint8mf2_t v1 = __riscv_vreinterpret_v_u8mf2_i8mf2(x_l); - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p0, p2)); - } + vint16m1_t vec_mul1 = __riscv_vwmul_vv_i16m1(v0, y0, vl); + vint16m1_t vec_mul2 = __riscv_vwmul_vv_i16m1(v1, y1, vl); - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + vint32m1_t vec_zero = __riscv_vmv_v_x_i32m1(0, vl); - } + vint32m1_t vs1 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul1, vec_zero, vl); + vint32m1_t vs2 = __riscv_vwredsum_vs_i16m1_i32m1(vec_mul2, vs1, vl); - *s = hsum_float_8(acc); + int sumi = __riscv_vmv_x_s_i32m1_i32(vs2); -#elif defined __AVX__ + sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d))*sumi + GGML_FP16_TO_FP32(x[i].m)*GGML_FP16_TO_FP32(y[i].s); + } - const __m128i m3 = _mm_set1_epi8(0x3); - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i m2 = _mm_set1_epi8(0x2); + *s = sumf; +#else + // scalar + float sumf = 0.0; - __m256 acc = _mm256_setzero_ps(); + for (int i = 0; i < nb; i++) { + uint32_t qh; + memcpy(&qh, x[i].qh, sizeof(qh)); - for (int i = 0; i < nb; ++i) { + int sumi = 0; - const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + for (int j = 0; j < qk/2; ++j) { + const uint8_t xh_0 = ((qh >> (j + 0)) << 4) & 0x10; + const uint8_t xh_1 = ((qh >> (j + 12)) ) & 0x10; - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; + const int32_t x0 = (x[i].qs[j] & 0xF) | xh_0; + const int32_t x1 = (x[i].qs[j] >> 4) | xh_1; - // load mins and scales from block_q2_K.scales[QK_K/16] - const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); - const __m128i scales16 = _mm_and_si128(mins_and_scales, m4); - const __m128i mins16 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); - const __m128i mins_0 = _mm_cvtepi8_epi16(mins16); - const __m128i mins_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(mins16, mins16)); + sumi += (x0 * y[i].qs[j]) + (x1 * y[i].qs[j + qk/2]); + } - // summs = y[i].bsums * (x[i].scales >> 4) in 16bits*8*2 to 32bits*4*2 - const __m128i summs_0 = _mm_madd_epi16(mins_0, _mm_loadu_si128((const __m128i*)&y[i].bsums[0])); - const __m128i summs_1 = _mm_madd_epi16(mins_1, _mm_loadu_si128((const __m128i*)&y[i].bsums[8])); + sumf += (GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d))*sumi + GGML_FP16_TO_FP32(x[i].m)*GGML_FP16_TO_FP32(y[i].s); + } - // sumf += -dmin * summs in 32bits*8 - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(MM256_SET_M128I(summs_1, summs_0))), acc); + *s = sumf; +#endif +} - const __m128i scales_0 = _mm_cvtepi8_epi16(scales16); - const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales16, scales16)); - const __m128i scales[2] = { scales_0, scales_1 }; +void ggml_vec_dot_q8_0_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + const int qk = QK8_0; + const int nb = n / qk; - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); + assert(n % qk == 0); +#if defined(__ARM_FEATURE_MATMUL_INT8) + assert((nrc == 2) || (nrc == 1)); +#else + assert(nrc == 1); +#endif + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - for (int j = 0; j < QK_K/128; ++j) { + const block_q8_0 * restrict x = vx; + const block_q8_0 * restrict y = vy; - // load Q8 quants int8*16*8 from block_q8_K.qs[QK_K] - const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; +#if defined(__ARM_FEATURE_MATMUL_INT8) + if (nrc == 2) { + const block_q8_0 * restrict vx0 = vx; + const block_q8_0 * restrict vx1 = vx + bx; + const block_q8_0 * restrict vy0 = vy; + const block_q8_0 * restrict vy1 = vy + by; - // load 2bits*16*8 from block_q2_K.qs[QK_K/4] - __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; - const __m128i q2_0 = _mm_and_si128(q2bits, m3); - const __m128i q2_2 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); - const __m128i q2_4 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); - const __m128i q2_6 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); - q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; - const __m128i q2_1 = _mm_and_si128(q2bits, m3); - const __m128i q2_3 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); - const __m128i q2_5 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); - const __m128i q2_7 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); + float32x4_t sumv0 = vdupq_n_f32(0.0f); - // isuml = q8[l] * ((q2[l] >> shift) & 3) in 8bits*16*8 to 16bits*8*8 - __m128i p0 = _mm_maddubs_epi16(q2_0, q8_0); - __m128i p1 = _mm_maddubs_epi16(q2_1, q8_1); - __m128i p2 = _mm_maddubs_epi16(q2_2, q8_2); - __m128i p3 = _mm_maddubs_epi16(q2_3, q8_3); - __m128i p4 = _mm_maddubs_epi16(q2_4, q8_4); - __m128i p5 = _mm_maddubs_epi16(q2_5, q8_5); - __m128i p6 = _mm_maddubs_epi16(q2_6, q8_6); - __m128i p7 = _mm_maddubs_epi16(q2_7, q8_7); + for (int i = 0; i < nb; i++) { + const block_q8_0 * restrict b_x0 = &vx0[i]; + const block_q8_0 * restrict b_y0 = &vy0[i]; - // isum += (x[i].scales[is++] & 0xF) * isuml in 16bits*8*8 to 32bits*4*8 - __m128i shuffle = _mm_set1_epi16(0x0100); - p0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p0); - shuffle = _mm_add_epi16(shuffle, m2); - p1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p1); - shuffle = _mm_add_epi16(shuffle, m2); - p2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p2); - shuffle = _mm_add_epi16(shuffle, m2); - p3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p3); - shuffle = _mm_add_epi16(shuffle, m2); - p4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p4); - shuffle = _mm_add_epi16(shuffle, m2); - p5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p5); - shuffle = _mm_add_epi16(shuffle, m2); - p6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p6); - shuffle = _mm_add_epi16(shuffle, m2); - p7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p7); + const block_q8_0 * restrict b_x1 = &vx1[i]; + const block_q8_0 * restrict b_y1 = &vy1[i]; - p0 = _mm_add_epi32(p0, p1); - p2 = _mm_add_epi32(p2, p3); - p4 = _mm_add_epi32(p4, p5); - p6 = _mm_add_epi32(p6, p7); + const int8x16_t x0_l = vld1q_s8(b_x0->qs); + const int8x16_t x0_h = vld1q_s8(b_x0->qs + 16); + const int8x16_t x1_l = vld1q_s8(b_x1->qs); + const int8x16_t x1_h = vld1q_s8(b_x1->qs + 16); - // isum in 32bits*4*2 - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p0, p2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p4, p6)); - } + // load y + const int8x16_t y0_l = vld1q_s8(b_y0->qs); + const int8x16_t y0_h = vld1q_s8(b_y0->qs + 16); + const int8x16_t y1_l = vld1q_s8(b_y1->qs); + const int8x16_t y1_h = vld1q_s8(b_y1->qs + 16); - // sumf += dall * isum - dmin * summs in 32bits - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dall), _mm256_cvtepi32_ps(sumi)), acc); - } + float32x4_t scale = {GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x0->d)*GGML_FP16_TO_FP32(b_y1->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y0->d), + GGML_FP16_TO_FP32(b_x1->d)*GGML_FP16_TO_FP32(b_y1->d)}; - *s = hsum_float_8(acc); + int8x16_t l0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); + int8x16_t l1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_l), vreinterpretq_s64_s8(x1_l))); -#elif defined __riscv_v_intrinsic + int8x16_t l2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); + int8x16_t l3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(x0_h), vreinterpretq_s64_s8(x1_h))); - float sumf = 0; - uint8_t temp_01[32] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; + int8x16_t r0 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); + int8x16_t r1 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_l), vreinterpretq_s64_s8(y1_l))); - for (int i = 0; i < nb; ++i) { + int8x16_t r2 = vreinterpretq_s8_s64(vzip1q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); + int8x16_t r3 = vreinterpretq_s8_s64(vzip2q_s64(vreinterpretq_s64_s8(y0_h), vreinterpretq_s64_s8(y1_h))); - const uint8_t * q2 = x[i].qs; - const int8_t * q8 = y[i].qs; - const uint8_t * sc = x[i].scales; + sumv0 = vmlaq_f32(sumv0,(vcvtq_f32_s32(vmmlaq_s32((vmmlaq_s32((vmmlaq_s32((vmmlaq_s32(vdupq_n_s32(0), l0, r0)), + l1, r1)), l2, r2)), l3, r3))), scale); + } + float32x4_t sumv1 = vextq_f32(sumv0, sumv0, 2); + float32x4_t sumv2 = vzip1q_f32(sumv0, sumv1); - const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + vst1_f32(s, vget_low_f32(sumv2)); + vst1_f32(s + bs, vget_high_f32(sumv2)); + return; + } +#endif +#if defined(__ARM_NEON) + float32x4_t sumv0 = vdupq_n_f32(0.0f); + float32x4_t sumv1 = vdupq_n_f32(0.0f); - size_t vl = 16; + assert(nb % 2 == 0); // TODO: handle odd nb - vuint8m1_t scales = __riscv_vle8_v_u8m1(sc, vl); - vuint8m1_t aux = __riscv_vand_vx_u8m1(scales, 0x0F, vl); + for (int i = 0; i < nb; i += 2) { + const block_q8_0 * restrict x0 = &x[i + 0]; + const block_q8_0 * restrict x1 = &x[i + 1]; + const block_q8_0 * restrict y0 = &y[i + 0]; + const block_q8_0 * restrict y1 = &y[i + 1]; - vint16m1_t q8sums = __riscv_vle16_v_i16m1(y[i].bsums, vl); + const int8x16_t x0_0 = vld1q_s8(x0->qs); + const int8x16_t x0_1 = vld1q_s8(x0->qs + 16); + const int8x16_t x1_0 = vld1q_s8(x1->qs); + const int8x16_t x1_1 = vld1q_s8(x1->qs + 16); - vuint8mf2_t scales_2 = __riscv_vle8_v_u8mf2(sc, vl); - vuint8mf2_t mins8 = __riscv_vsrl_vx_u8mf2(scales_2, 0x4, vl); - vint16m1_t mins = __riscv_vreinterpret_v_u16m1_i16m1(__riscv_vzext_vf2_u16m1(mins8, vl)); - vint32m2_t prod = __riscv_vwmul_vv_i32m2(q8sums, mins, vl); - vint32m1_t vsums = __riscv_vredsum_vs_i32m2_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); + // load y + const int8x16_t y0_0 = vld1q_s8(y0->qs); + const int8x16_t y0_1 = vld1q_s8(y0->qs + 16); + const int8x16_t y1_0 = vld1q_s8(y1->qs); + const int8x16_t y1_1 = vld1q_s8(y1->qs + 16); - sumf += dmin * __riscv_vmv_x_s_i32m1_i32(vsums); - - vl = 32; - - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vuint8m1_t v_b = __riscv_vle8_v_u8m1(temp_01, vl); - - uint8_t is=0; - int isum=0; - - for (int j = 0; j < QK_K/128; ++j) { - // load Q2 - vuint8m1_t q2_x = __riscv_vle8_v_u8m1(q2, vl); + sumv0 = vmlaq_n_f32(sumv0, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), x0_0, y0_0), + ggml_vdotq_s32(vdupq_n_s32(0), x0_1, y0_1))), GGML_FP16_TO_FP32(x0->d)*GGML_FP16_TO_FP32(y0->d)); - vuint8m1_t q2_0 = __riscv_vand_vx_u8m1(q2_x, 0x03, vl); - vuint8m1_t q2_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x2, vl), 0x03 , vl); - vuint8m1_t q2_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x4, vl), 0x03 , vl); - vuint8m1_t q2_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x6, vl), 0x03 , vl); + sumv1 = vmlaq_n_f32(sumv1, vcvtq_f32_s32(vaddq_s32( + ggml_vdotq_s32(vdupq_n_s32(0), x1_0, y1_0), + ggml_vdotq_s32(vdupq_n_s32(0), x1_1, y1_1))), GGML_FP16_TO_FP32(x1->d)*GGML_FP16_TO_FP32(y1->d)); + } - // duplicate scale elements for product - vuint8m1_t sc0 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 0+is, vl), vl); - vuint8m1_t sc1 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 2+is, vl), vl); - vuint8m1_t sc2 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 4+is, vl), vl); - vuint8m1_t sc3 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 6+is, vl), vl); + *s = vaddvq_f32(sumv0) + vaddvq_f32(sumv1); +#elif defined(__AVX2__) || defined(__AVX__) + // Initialize accumulator with zeros + __m256 acc = _mm256_setzero_ps(); - vint16m2_t p0 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_0, sc0, vl)); - vint16m2_t p1 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_1, sc1, vl)); - vint16m2_t p2 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_2, sc2, vl)); - vint16m2_t p3 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_3, sc3, vl)); + // Main loop + for (int i = 0; i < nb; ++i) { + // Compute combined scale for the block + const __m256 d = _mm256_set1_ps(GGML_FP16_TO_FP32(x[i].d) * GGML_FP16_TO_FP32(y[i].d)); + __m256i qx = _mm256_loadu_si256((const __m256i *)x[i].qs); + __m256i qy = _mm256_loadu_si256((const __m256i *)y[i].qs); - // load Q8 - vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); - vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); - vint8m1_t q8_2 = __riscv_vle8_v_i8m1(q8+64, vl); - vint8m1_t q8_3 = __riscv_vle8_v_i8m1(q8+96, vl); + const __m256 q = mul_sum_i8_pairs_float(qx, qy); - vint32m4_t s0 = __riscv_vwmul_vv_i32m4(p0, __riscv_vwcvt_x_x_v_i16m2(q8_0, vl), vl); - vint32m4_t s1 = __riscv_vwmul_vv_i32m4(p1, __riscv_vwcvt_x_x_v_i16m2(q8_1, vl), vl); - vint32m4_t s2 = __riscv_vwmul_vv_i32m4(p2, __riscv_vwcvt_x_x_v_i16m2(q8_2, vl), vl); - vint32m4_t s3 = __riscv_vwmul_vv_i32m4(p3, __riscv_vwcvt_x_x_v_i16m2(q8_3, vl), vl); + // Multiply q with scale and accumulate +#if defined(__AVX2__) + acc = _mm256_fmadd_ps( d, q, acc ); +#else + acc = _mm256_add_ps( _mm256_mul_ps( d, q ), acc ); +#endif + } - vint32m1_t isum0 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s0, s1, vl), vzero, vl); - vint32m1_t isum1 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s2, s3, vl), isum0, vl); + *s = hsum_float_8(acc); +#elif defined(__riscv_v_intrinsic) + float sumf = 0.0; + size_t vl = __riscv_vsetvl_e8m1(qk); - isum += __riscv_vmv_x_s_i32m1_i32(isum1); + for (int i = 0; i < nb; i++) { + // load elements + vint8m1_t bx_0 = __riscv_vle8_v_i8m1(x[i].qs, vl); + vint8m1_t by_0 = __riscv_vle8_v_i8m1(y[i].qs, vl); - q2+=32; q8+=128; is=8; + vint16m2_t vw_mul = __riscv_vwmul_vv_i16m2(bx_0, by_0, vl); - } + vint32m1_t v_zero = __riscv_vmv_v_x_i32m1(0, vl); + vint32m1_t v_sum = __riscv_vwredsum_vs_i16m2_i32m1(vw_mul, v_zero, vl); - sumf += dall * isum; + int sumi = __riscv_vmv_x_s_i32m1_i32(v_sum); + sumf += sumi*(GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)); } *s = sumf; - #else + // scalar + float sumf = 0.0; - float sumf = 0; - - for (int i = 0; i < nb; ++i) { - - const uint8_t * q2 = x[i].qs; - const int8_t * q8 = y[i].qs; - const uint8_t * sc = x[i].scales; + for (int i = 0; i < nb; i++) { + int sumi = 0; - int summs = 0; - for (int j = 0; j < 16; ++j) { - summs += y[i].bsums[j] * (sc[j] >> 4); + for (int j = 0; j < qk; j++) { + sumi += x[i].qs[j]*y[i].qs[j]; } - const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - int isum = 0; - int is = 0; - int d; - for (int k = 0; k < QK_K/128; ++k) { - int shift = 0; - for (int j = 0; j < 4; ++j) { - d = sc[is++] & 0xF; - int isuml = 0; - for (int l = 0; l < 16; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); - isum += d * isuml; - d = sc[is++] & 0xF; - isuml = 0; - for (int l = 16; l < 32; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); - isum += d * isuml; - shift += 2; - q8 += 32; - } - q2 += 32; - } - sumf += dall * isum - dmin * summs; + sumf += sumi*(GGML_FP16_TO_FP32(x[i].d)*GGML_FP16_TO_FP32(y[i].d)); } + *s = sumf; #endif } -#else - -void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { +#if QK_K == 256 +void ggml_vec_dot_q2_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); const block_q2_K * restrict x = vx; const block_q8_K * restrict y = vy; @@ -3943,66 +5041,69 @@ void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restri const int nb = n / QK_K; #ifdef __ARM_NEON - const uint8x16_t m3 = vdupq_n_u8(0x3); -#if defined(__ARM_FEATURE_DOTPROD) - const int32x4_t vzero = vdupq_n_s32(0); -#endif + const uint8x16_t m4 = vdupq_n_u8(0xF); - int8x16x4_t q2bytes; + const int32x4_t vzero = vdupq_n_s32(0); - uint32_t aux32[2]; - const uint8_t * scales = (const uint8_t *)aux32; + ggml_int8x16x2_t q2bytes; + uint8_t aux[16]; float sum = 0; for (int i = 0; i < nb; ++i) { - - const float d = y[i].d * (float)x[i].d; - const float dmin = -y[i].d * (float)x[i].dmin; + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); const uint8_t * restrict q2 = x[i].qs; const int8_t * restrict q8 = y[i].qs; - const uint32_t * restrict sc = (const uint32_t *)x[i].scales; + const uint8_t * restrict sc = x[i].scales; - aux32[0] = sc[0] & 0x0f0f0f0f; - aux32[1] = (sc[0] >> 4) & 0x0f0f0f0f; + const uint8x16_t mins_and_scales = vld1q_u8(sc); + const uint8x16_t scales = vandq_u8(mins_and_scales, m4); + vst1q_u8(aux, scales); - sum += dmin * (scales[4] * y[i].bsums[0] + scales[5] * y[i].bsums[1] + scales[6] * y[i].bsums[2] + scales[7] * y[i].bsums[3]); + const uint8x16_t mins = vshrq_n_u8(mins_and_scales, 4); + const ggml_int16x8x2_t q8sums = ggml_vld1q_s16_x2(y[i].bsums); + const ggml_int16x8x2_t mins16 = {{vreinterpretq_s16_u16(vmovl_u8(vget_low_u8(mins))), vreinterpretq_s16_u16(vmovl_u8(vget_high_u8(mins)))}}; + const int32x4_t s0 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[0]), vget_low_s16 (q8sums.val[0])), + vmull_s16(vget_high_s16(mins16.val[0]), vget_high_s16(q8sums.val[0]))); + const int32x4_t s1 = vaddq_s32(vmull_s16(vget_low_s16 (mins16.val[1]), vget_low_s16 (q8sums.val[1])), + vmull_s16(vget_high_s16(mins16.val[1]), vget_high_s16(q8sums.val[1]))); + sum += dmin * vaddvq_s32(vaddq_s32(s0, s1)); - int isum1 = 0, isum2 = 0; + int isum = 0; + int is = 0; - const uint8x16_t q2bits = vld1q_u8(q2); +// We use this macro instead of a function call because for some reason +// the code runs 2-3% slower, even if the function is declared inline +#define MULTIPLY_ACCUM_WITH_SCALE(index)\ + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[0], q8bytes.val[0])) * aux[is+(index)];\ + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[1], q8bytes.val[1])) * aux[is+1+(index)]; - const int8x16x4_t q8bytes = vld1q_s8_x4(q8); +#define SHIFT_MULTIPLY_ACCUM_WITH_SCALE(shift, index)\ + q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32;\ + q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[0], (shift)), m3));\ + q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits.val[1], (shift)), m3));\ + MULTIPLY_ACCUM_WITH_SCALE((index)); - q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(q2bits, m3)); - q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits, 2), m3)); - q2bytes.val[2] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits, 4), m3)); - q2bytes.val[3] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits, 6), m3)); + for (int j = 0; j < QK_K/128; ++j) { + const ggml_uint8x16x2_t q2bits = ggml_vld1q_u8_x2(q2); q2 += 32; -#if defined(__ARM_FEATURE_DOTPROD) - isum1 += vaddvq_s32(vdotq_s32(vzero, q2bytes.val[0], q8bytes.val[0])) * scales[0]; - isum2 += vaddvq_s32(vdotq_s32(vzero, q2bytes.val[1], q8bytes.val[1])) * scales[1]; - isum1 += vaddvq_s32(vdotq_s32(vzero, q2bytes.val[2], q8bytes.val[2])) * scales[2]; - isum2 += vaddvq_s32(vdotq_s32(vzero, q2bytes.val[3], q8bytes.val[3])) * scales[3]; -#else - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q2bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q2bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q2bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q2bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - isum1 += vaddvq_s16(p1) * scales[0]; - isum2 += vaddvq_s16(p2) * scales[1]; - - const int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q2bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q2bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - const int16x8_t p4 = vaddq_s16(vmull_s8(vget_low_s8 (q2bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q2bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - isum1 += vaddvq_s16(p3) * scales[2]; - isum2 += vaddvq_s16(p4) * scales[3]; -#endif - sum += d * (isum1 + isum2); + ggml_int8x16x2_t q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; + q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[0], m3)); + q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(q2bits.val[1], m3)); + + MULTIPLY_ACCUM_WITH_SCALE(0); + + SHIFT_MULTIPLY_ACCUM_WITH_SCALE(2, 2); + SHIFT_MULTIPLY_ACCUM_WITH_SCALE(4, 4); + SHIFT_MULTIPLY_ACCUM_WITH_SCALE(6, 6); + + is += 8; + } + sum += d * isum; } *s = sum; @@ -4010,17 +5111,10 @@ void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restri #elif defined __AVX2__ const __m256i m3 = _mm256_set1_epi8(3); + const __m128i m4 = _mm_set1_epi8(0xF); __m256 acc = _mm256_setzero_ps(); - uint32_t ud, um; - const uint8_t * restrict db = (const uint8_t *)&ud; - const uint8_t * restrict mb = (const uint8_t *)&um; - - float summs = 0; - - // TODO: optimize this - for (int i = 0; i < nb; ++i) { const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); @@ -4029,145 +5123,242 @@ void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restri const uint8_t * restrict q2 = x[i].qs; const int8_t * restrict q8 = y[i].qs; - const uint32_t * restrict sc = (const uint32_t *)x[i].scales; - ud = (sc[0] >> 0) & 0x0f0f0f0f; - um = (sc[0] >> 4) & 0x0f0f0f0f; + const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); + const __m128i scales8 = _mm_and_si128(mins_and_scales, m4); + const __m128i mins8 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); + const __m256i mins = _mm256_cvtepi8_epi16(mins8); + const __m256i prod = _mm256_madd_epi16(mins, _mm256_loadu_si256((const __m256i*)y[i].bsums)); - int32_t smin = mb[0] * y[i].bsums[0] + mb[1] * y[i].bsums[1] + mb[2] * y[i].bsums[2] + mb[3] * y[i].bsums[3]; - summs += dmin * smin; + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(prod), acc); - const __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); - const __m256i q2_0 = _mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q2bits, 2), q2bits), m3); - const __m256i q2_1 = _mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q2bits, 6), _mm_srli_epi16(q2bits, 4)), m3); + const __m256i all_scales = _mm256_cvtepi8_epi16(scales8); + const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); + const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); + const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + __m256i sumi = _mm256_setzero_si256(); - const __m256i p0 = _mm256_maddubs_epi16(q2_0, q8_0); - const __m256i p1 = _mm256_maddubs_epi16(q2_1, q8_1); + for (int j = 0; j < QK_K/128; ++j) { - const __m256i p_0 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p0, 0)); - const __m256i p_1 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p0, 1)); - const __m256i p_2 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p1, 0)); - const __m256i p_3 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p1, 1)); + const __m256i q2bits = _mm256_loadu_si256((const __m256i*)q2); q2 += 32; - acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[0]), _mm256_cvtepi32_ps(p_0), acc); - acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[1]), _mm256_cvtepi32_ps(p_1), acc); - acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[2]), _mm256_cvtepi32_ps(p_2), acc); - acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[3]), _mm256_cvtepi32_ps(p_3), acc); - } + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - *s = hsum_float_8(acc) + summs; + const __m256i q2_0 = _mm256_and_si256(q2bits, m3); + const __m256i q2_1 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 2), m3); + const __m256i q2_2 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 4), m3); + const __m256i q2_3 = _mm256_and_si256(_mm256_srli_epi16(q2bits, 6), m3); -#elif defined __AVX__ + __m256i p0 = _mm256_maddubs_epi16(q2_0, q8_0); + __m256i p1 = _mm256_maddubs_epi16(q2_1, q8_1); + __m256i p2 = _mm256_maddubs_epi16(q2_2, q8_2); + __m256i p3 = _mm256_maddubs_epi16(q2_3, q8_3); - const __m128i m3 = _mm_set1_epi8(3); + p0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(0)), p0); + p1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(1)), p1); + p2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(2)), p2); + p3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(3)), p3); - __m256 acc = _mm256_setzero_ps(); + p0 = _mm256_add_epi32(p0, p1); + p2 = _mm256_add_epi32(p2, p3); - uint32_t ud, um; - const uint8_t * restrict db = (const uint8_t *)&ud; - const uint8_t * restrict mb = (const uint8_t *)&um; + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p0, p2)); + } - float summs = 0; + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); - // TODO: optimize this + } + + *s = hsum_float_8(acc); + +#elif defined __AVX__ + + const __m128i m3 = _mm_set1_epi8(0x3); + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i m2 = _mm_set1_epi8(0x2); + + __m256 acc = _mm256_setzero_ps(); for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); const uint8_t * restrict q2 = x[i].qs; const int8_t * restrict q8 = y[i].qs; - const uint32_t * restrict sc = (const uint32_t *)x[i].scales; - ud = (sc[0] >> 0) & 0x0f0f0f0f; - um = (sc[0] >> 4) & 0x0f0f0f0f; + // load mins and scales from block_q2_K.scales[QK_K/16] + const __m128i mins_and_scales = _mm_loadu_si128((const __m128i*)x[i].scales); + const __m128i scales16 = _mm_and_si128(mins_and_scales, m4); + const __m128i mins16 = _mm_and_si128(_mm_srli_epi16(mins_and_scales, 4), m4); + const __m128i mins_0 = _mm_cvtepi8_epi16(mins16); + const __m128i mins_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(mins16, mins16)); - int32_t smin = mb[0] * y[i].bsums[0] + mb[1] * y[i].bsums[1] + mb[2] * y[i].bsums[2] + mb[3] * y[i].bsums[3]; - summs += dmin * smin; + // summs = y[i].bsums * (x[i].scales >> 4) in 16bits*8*2 to 32bits*4*2 + const __m128i summs_0 = _mm_madd_epi16(mins_0, _mm_loadu_si128((const __m128i*)&y[i].bsums[0])); + const __m128i summs_1 = _mm_madd_epi16(mins_1, _mm_loadu_si128((const __m128i*)&y[i].bsums[8])); - const __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); - const __m128i q2_0 = _mm_and_si128(q2bits, m3); - const __m128i q2_1 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); - const __m128i q2_2 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); - const __m128i q2_3 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); + // sumf += -dmin * summs in 32bits*8 + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dmin), _mm256_cvtepi32_ps(MM256_SET_M128I(summs_1, summs_0))), acc); - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + const __m128i scales_0 = _mm_cvtepi8_epi16(scales16); + const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales16, scales16)); + const __m128i scales[2] = { scales_0, scales_1 }; - const __m128i p0 = _mm_maddubs_epi16(q2_0, _mm256_extractf128_si256(q8_0, 0)); - const __m128i p1 = _mm_maddubs_epi16(q2_1, _mm256_extractf128_si256(q8_0, 1)); - const __m128i p2 = _mm_maddubs_epi16(q2_2, _mm256_extractf128_si256(q8_1, 0)); - const __m128i p3 = _mm_maddubs_epi16(q2_3, _mm256_extractf128_si256(q8_1, 1)); + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); - const __m256i p_0 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p0, p0)), _mm_cvtepi16_epi32(p0)); - const __m256i p_1 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p1, p1)), _mm_cvtepi16_epi32(p1)); - const __m256i p_2 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p2, p2)), _mm_cvtepi16_epi32(p2)); - const __m256i p_3 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p3, p3)), _mm_cvtepi16_epi32(p3)); + for (int j = 0; j < QK_K/128; ++j) { - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[0]), _mm256_cvtepi32_ps(p_0)), acc); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[1]), _mm256_cvtepi32_ps(p_1)), acc); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[2]), _mm256_cvtepi32_ps(p_2)), acc); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[3]), _mm256_cvtepi32_ps(p_3)), acc); + // load Q8 quants int8*16*8 from block_q8_K.qs[QK_K] + const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + + // load 2bits*16*8 from block_q2_K.qs[QK_K/4] + __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; + const __m128i q2_0 = _mm_and_si128(q2bits, m3); + const __m128i q2_2 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); + const __m128i q2_4 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); + const __m128i q2_6 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); + q2bits = _mm_loadu_si128((const __m128i*)q2); q2 += 16; + const __m128i q2_1 = _mm_and_si128(q2bits, m3); + const __m128i q2_3 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); + const __m128i q2_5 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); + const __m128i q2_7 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); + + // isuml = q8[l] * ((q2[l] >> shift) & 3) in 8bits*16*8 to 16bits*8*8 + __m128i p0 = _mm_maddubs_epi16(q2_0, q8_0); + __m128i p1 = _mm_maddubs_epi16(q2_1, q8_1); + __m128i p2 = _mm_maddubs_epi16(q2_2, q8_2); + __m128i p3 = _mm_maddubs_epi16(q2_3, q8_3); + __m128i p4 = _mm_maddubs_epi16(q2_4, q8_4); + __m128i p5 = _mm_maddubs_epi16(q2_5, q8_5); + __m128i p6 = _mm_maddubs_epi16(q2_6, q8_6); + __m128i p7 = _mm_maddubs_epi16(q2_7, q8_7); + + // isum += (x[i].scales[is++] & 0xF) * isuml in 16bits*8*8 to 32bits*4*8 + __m128i shuffle = _mm_set1_epi16(0x0100); + p0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p0); + shuffle = _mm_add_epi16(shuffle, m2); + p1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p1); + shuffle = _mm_add_epi16(shuffle, m2); + p2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p2); + shuffle = _mm_add_epi16(shuffle, m2); + p3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p3); + shuffle = _mm_add_epi16(shuffle, m2); + p4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p4); + shuffle = _mm_add_epi16(shuffle, m2); + p5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p5); + shuffle = _mm_add_epi16(shuffle, m2); + p6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p6); + shuffle = _mm_add_epi16(shuffle, m2); + p7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p7); + + p0 = _mm_add_epi32(p0, p1); + p2 = _mm_add_epi32(p2, p3); + p4 = _mm_add_epi32(p4, p5); + p6 = _mm_add_epi32(p6, p7); + + // isum in 32bits*4*2 + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p0, p2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p4, p6)); + } + + // sumf += dall * isum - dmin * summs in 32bits + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&dall), _mm256_cvtepi32_ps(sumi)), acc); } - *s = hsum_float_8(acc) + summs; + *s = hsum_float_8(acc); #elif defined __riscv_v_intrinsic - uint32_t aux32[2]; - const uint8_t * scales = (const uint8_t *)aux32; - float sumf = 0; + uint8_t temp_01[32] = {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, + 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1}; for (int i = 0; i < nb; ++i) { - const float d = y[i].d * (float)x[i].d; - const float dmin = -y[i].d * (float)x[i].dmin; + const uint8_t * q2 = x[i].qs; + const int8_t * q8 = y[i].qs; + const uint8_t * sc = x[i].scales; - const uint8_t * restrict q2 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - const uint32_t * restrict sc = (const uint32_t *)x[i].scales; + const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - aux32[0] = sc[0] & 0x0f0f0f0f; - aux32[1] = (sc[0] >> 4) & 0x0f0f0f0f; + size_t vl = 16; - sumf += dmin * (scales[4] * y[i].bsums[0] + scales[5] * y[i].bsums[1] + scales[6] * y[i].bsums[2] + scales[7] * y[i].bsums[3]); + vuint8m1_t scales = __riscv_vle8_v_u8m1(sc, vl); + vuint8m1_t aux = __riscv_vand_vx_u8m1(scales, 0x0F, vl); - int isum1 = 0; - int isum2 = 0; + vint16m1_t q8sums = __riscv_vle16_v_i16m1(y[i].bsums, vl); - size_t vl = 16; + vuint8mf2_t scales_2 = __riscv_vle8_v_u8mf2(sc, vl); + vuint8mf2_t mins8 = __riscv_vsrl_vx_u8mf2(scales_2, 0x4, vl); + vint16m1_t mins = __riscv_vreinterpret_v_u16m1_i16m1(__riscv_vzext_vf2_u16m1(mins8, vl)); + vint32m2_t prod = __riscv_vwmul_vv_i32m2(q8sums, mins, vl); + vint32m1_t vsums = __riscv_vredsum_vs_i32m2_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); - vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); + sumf += dmin * __riscv_vmv_x_s_i32m1_i32(vsums); - // load Q2 - vuint8mf2_t q2_x = __riscv_vle8_v_u8mf2(q2, vl); + vl = 32; - vint8mf2_t q2_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(q2_x, 0x03, vl)); - vint8mf2_t q2_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q2_x, 0x2, vl), 0x03 , vl)); - vint8mf2_t q2_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q2_x, 0x4, vl), 0x03 , vl)); - vint8mf2_t q2_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q2_x, 0x6, vl), 0x03 , vl)); + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + vuint8m1_t v_b = __riscv_vle8_v_u8m1(temp_01, vl); - // load Q8, and take product with Q2 - vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q2_0, __riscv_vle8_v_i8mf2(q8, vl), vl); - vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q2_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); - vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q2_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); - vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q2_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); + uint8_t is=0; + int isum=0; - vint16m1_t vs_0 = __riscv_vredsum_vs_i16m1_i16m1(p0, vzero, vl); - vint16m1_t vs_1 = __riscv_vredsum_vs_i16m1_i16m1(p1, vzero, vl); - vint16m1_t vs_2 = __riscv_vredsum_vs_i16m1_i16m1(p2, vzero, vl); - vint16m1_t vs_3 = __riscv_vredsum_vs_i16m1_i16m1(p3, vzero, vl); + for (int j = 0; j < QK_K/128; ++j) { + // load Q2 + vuint8m1_t q2_x = __riscv_vle8_v_u8m1(q2, vl); - isum1 += __riscv_vmv_x_s_i16m1_i16(vs_0) * scales[0]; - isum2 += __riscv_vmv_x_s_i16m1_i16(vs_1) * scales[1]; - isum1 += __riscv_vmv_x_s_i16m1_i16(vs_2) * scales[2]; - isum2 += __riscv_vmv_x_s_i16m1_i16(vs_3) * scales[3]; + vuint8m1_t q2_0 = __riscv_vand_vx_u8m1(q2_x, 0x03, vl); + vuint8m1_t q2_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x2, vl), 0x03 , vl); + vuint8m1_t q2_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x4, vl), 0x03 , vl); + vuint8m1_t q2_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q2_x, 0x6, vl), 0x03 , vl); - sumf += d * (isum1 + isum2); + // duplicate scale elements for product + vuint8m1_t sc0 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 0+is, vl), vl); + vuint8m1_t sc1 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 2+is, vl), vl); + vuint8m1_t sc2 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 4+is, vl), vl); + vuint8m1_t sc3 = __riscv_vrgather_vv_u8m1(aux, __riscv_vadd_vx_u8m1(v_b, 6+is, vl), vl); + + vint16m2_t p0 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_0, sc0, vl)); + vint16m2_t p1 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_1, sc1, vl)); + vint16m2_t p2 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_2, sc2, vl)); + vint16m2_t p3 = __riscv_vreinterpret_v_u16m2_i16m2(__riscv_vwmulu_vv_u16m2(q2_3, sc3, vl)); + + // load Q8 + vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); + vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); + vint8m1_t q8_2 = __riscv_vle8_v_i8m1(q8+64, vl); + vint8m1_t q8_3 = __riscv_vle8_v_i8m1(q8+96, vl); + + vint32m4_t s0 = __riscv_vwmul_vv_i32m4(p0, __riscv_vwcvt_x_x_v_i16m2(q8_0, vl), vl); + vint32m4_t s1 = __riscv_vwmul_vv_i32m4(p1, __riscv_vwcvt_x_x_v_i16m2(q8_1, vl), vl); + vint32m4_t s2 = __riscv_vwmul_vv_i32m4(p2, __riscv_vwcvt_x_x_v_i16m2(q8_2, vl), vl); + vint32m4_t s3 = __riscv_vwmul_vv_i32m4(p3, __riscv_vwcvt_x_x_v_i16m2(q8_3, vl), vl); + + vint32m1_t isum0 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s0, s1, vl), vzero, vl); + vint32m1_t isum1 = __riscv_vredsum_vs_i32m4_i32m1(__riscv_vadd_vv_i32m4(s2, s3, vl), isum0, vl); + + isum += __riscv_vmv_x_s_i32m1_i32(isum1); + + q2+=32; q8+=128; is=8; + + } + + sumf += dall * isum; } @@ -4177,8 +5368,6 @@ void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restri float sumf = 0; - int isum[4]; - for (int i = 0; i < nb; ++i) { const uint8_t * q2 = x[i].qs; @@ -4186,156 +5375,95 @@ void ggml_vec_dot_q2_K_q8_K(const int n, float * restrict s, const void * restri const uint8_t * sc = x[i].scales; int summs = 0; - for (int j = 0; j < QK_K/16; ++j) { + for (int j = 0; j < 16; ++j) { summs += y[i].bsums[j] * (sc[j] >> 4); } const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - isum[0] = isum[1] = isum[2] = isum[3] = 0; - for (int l = 0; l < 16; ++l) { - isum[0] += q8[l+ 0] * ((q2[l] >> 0) & 3); - isum[1] += q8[l+16] * ((q2[l] >> 2) & 3); - isum[2] += q8[l+32] * ((q2[l] >> 4) & 3); - isum[3] += q8[l+48] * ((q2[l] >> 6) & 3); - } - for (int l = 0; l < 4; ++l) { - isum[l] *= (sc[l] & 0xF); + int isum = 0; + int is = 0; + int d; + for (int k = 0; k < QK_K/128; ++k) { + int shift = 0; + for (int j = 0; j < 4; ++j) { + d = sc[is++] & 0xF; + int isuml = 0; + for (int l = 0; l < 16; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); + isum += d * isuml; + d = sc[is++] & 0xF; + isuml = 0; + for (int l = 16; l < 32; ++l) isuml += q8[l] * ((q2[l] >> shift) & 3); + isum += d * isuml; + shift += 2; + q8 += 32; + } + q2 += 32; } - sumf += dall * (isum[0] + isum[1] + isum[2] + isum[3]) - dmin * summs; + sumf += dall * isum - dmin * summs; } *s = sumf; #endif } -#endif -#if QK_K == 256 -void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - assert(n % QK_K == 0); +#else - const uint32_t kmask1 = 0x03030303; - const uint32_t kmask2 = 0x0f0f0f0f; +void ggml_vec_dot_q2_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q3_K * restrict x = vx; + const block_q2_K * restrict x = vx; const block_q8_K * restrict y = vy; const int nb = n / QK_K; #ifdef __ARM_NEON + const uint8x16_t m3 = vdupq_n_u8(0x3); - uint32_t aux[3]; - uint32_t utmp[4]; - - const uint8x16_t m3b = vdupq_n_u8(0x3); -#ifdef __ARM_FEATURE_DOTPROD - const int32x4_t vzero = vdupq_n_s32(0); -#endif + const int32x4_t vzero = vdupq_n_s32(0); - const uint8x16_t m0 = vdupq_n_u8(1); - const uint8x16_t m1 = vshlq_n_u8(m0, 1); - const uint8x16_t m2 = vshlq_n_u8(m0, 2); - const uint8x16_t m3 = vshlq_n_u8(m0, 3); - const int8_t m32 = 32; + ggml_int8x16x4_t q2bytes; - int8x16x4_t q3bytes; + uint32_t aux32[2]; + const uint8_t * scales = (const uint8_t *)aux32; float sum = 0; for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict qh = x[i].hmask; + const uint8_t * restrict q2 = x[i].qs; const int8_t * restrict q8 = y[i].qs; + const uint32_t * restrict sc = (const uint32_t *)x[i].scales; - uint8x16x2_t qhbits = vld1q_u8_x2(qh); - - uint8x16x4_t q3h; - - int32_t isum = 0; - - // Set up scales - memcpy(aux, x[i].scales, 12); - utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); - utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); - utmp[1] = (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4); - utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); - - int8_t * scale = (int8_t *)utmp; - for (int j = 0; j < 16; ++j) scale[j] -= m32; - - for (int j = 0; j < QK_K/128; ++j) { - - const uint8x16x2_t q3bits = vld1q_u8_x2(q3); q3 += 32; - const int8x16x4_t q8bytes_1 = vld1q_s8_x4(q8); q8 += 64; - const int8x16x4_t q8bytes_2 = vld1q_s8_x4(q8); q8 += 64; - - q3h.val[0] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[0]), 2); - q3h.val[1] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[1]), 2); - q3h.val[2] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[0]), 1); - q3h.val[3] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[1]), 1); - - q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[0], m3b)), vreinterpretq_s8_u8(q3h.val[0])); - q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[1], m3b)), vreinterpretq_s8_u8(q3h.val[1])); - q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 2), m3b)), vreinterpretq_s8_u8(q3h.val[2])); - q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 2), m3b)), vreinterpretq_s8_u8(q3h.val[3])); + aux32[0] = sc[0] & 0x0f0f0f0f; + aux32[1] = (sc[0] >> 4) & 0x0f0f0f0f; -#if defined(__ARM_FEATURE_DOTPROD) - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[0], q8bytes_1.val[0])) * scale[0]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[1], q8bytes_1.val[1])) * scale[1]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[2], q8bytes_1.val[2])) * scale[2]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[3], q8bytes_1.val[3])) * scale[3]; -#else - int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[0]), vget_low_s8 (q8bytes_1.val[0])), - vmull_s8(vget_high_s8(q3bytes.val[0]), vget_high_s8(q8bytes_1.val[0]))); - int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[1]), vget_low_s8 (q8bytes_1.val[1])), - vmull_s8(vget_high_s8(q3bytes.val[1]), vget_high_s8(q8bytes_1.val[1]))); - int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[2]), vget_low_s8 (q8bytes_1.val[2])), - vmull_s8(vget_high_s8(q3bytes.val[2]), vget_high_s8(q8bytes_1.val[2]))); - int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[3]), vget_low_s8 (q8bytes_1.val[3])), - vmull_s8(vget_high_s8(q3bytes.val[3]), vget_high_s8(q8bytes_1.val[3]))); - isum += vaddvq_s16(p0) * scale[0] + vaddvq_s16(p1) * scale[1] + vaddvq_s16(p2) * scale[2] + vaddvq_s16(p3) * scale[3]; -#endif - scale += 4; + sum += dmin * (scales[4] * y[i].bsums[0] + scales[5] * y[i].bsums[1] + scales[6] * y[i].bsums[2] + scales[7] * y[i].bsums[3]); - q3h.val[0] = vbicq_u8(m2, qhbits.val[0]); - q3h.val[1] = vbicq_u8(m2, qhbits.val[1]); - q3h.val[2] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[0]), 1); - q3h.val[3] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[1]), 1); + int isum1 = 0, isum2 = 0; - q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 4), m3b)), vreinterpretq_s8_u8(q3h.val[0])); - q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 4), m3b)), vreinterpretq_s8_u8(q3h.val[1])); - q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 6), m3b)), vreinterpretq_s8_u8(q3h.val[2])); - q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 6), m3b)), vreinterpretq_s8_u8(q3h.val[3])); + const uint8x16_t q2bits = vld1q_u8(q2); -#if defined(__ARM_FEATURE_DOTPROD) - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[0], q8bytes_2.val[0])) * scale[0]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[1], q8bytes_2.val[1])) * scale[1]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[2], q8bytes_2.val[2])) * scale[2]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[3], q8bytes_2.val[3])) * scale[3]; -#else - p0 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[0]), vget_low_s8 (q8bytes_2.val[0])), - vmull_s8(vget_high_s8(q3bytes.val[0]), vget_high_s8(q8bytes_2.val[0]))); - p1 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[1]), vget_low_s8 (q8bytes_2.val[1])), - vmull_s8(vget_high_s8(q3bytes.val[1]), vget_high_s8(q8bytes_2.val[1]))); - p2 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[2]), vget_low_s8 (q8bytes_2.val[2])), - vmull_s8(vget_high_s8(q3bytes.val[2]), vget_high_s8(q8bytes_2.val[2]))); - p3 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[3]), vget_low_s8 (q8bytes_2.val[3])), - vmull_s8(vget_high_s8(q3bytes.val[3]), vget_high_s8(q8bytes_2.val[3]))); - isum += vaddvq_s16(p0) * scale[0] + vaddvq_s16(p1) * scale[1] + vaddvq_s16(p2) * scale[2] + vaddvq_s16(p3) * scale[3]; -#endif - scale += 4; + const ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); - if (j == 0) { - qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 4); - qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 4); - } + q2bytes.val[0] = vreinterpretq_s8_u8(vandq_u8(q2bits, m3)); + q2bytes.val[1] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits, 2), m3)); + q2bytes.val[2] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits, 4), m3)); + q2bytes.val[3] = vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q2bits, 6), m3)); - } - sum += d * isum; + isum1 += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[0], q8bytes.val[0])) * scales[0]; + isum2 += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[1], q8bytes.val[1])) * scales[1]; + isum1 += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[2], q8bytes.val[2])) * scales[2]; + isum2 += vaddvq_s32(ggml_vdotq_s32(vzero, q2bytes.val[3], q8bytes.val[3])) * scales[3]; + sum += d * (isum1 + isum2); } *s = sum; @@ -4343,261 +5471,256 @@ void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restri #elif defined __AVX2__ const __m256i m3 = _mm256_set1_epi8(3); - const __m256i mone = _mm256_set1_epi8(1); - const __m128i m32 = _mm_set1_epi8(32); __m256 acc = _mm256_setzero_ps(); - uint32_t aux[3]; + uint32_t ud, um; + const uint8_t * restrict db = (const uint8_t *)&ud; + const uint8_t * restrict mb = (const uint8_t *)&um; + + float summs = 0; + + // TODO: optimize this for (int i = 0; i < nb; ++i) { const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict q2 = x[i].qs; const int8_t * restrict q8 = y[i].qs; - // Set up scales - memcpy(aux, x[i].scales, 12); - __m128i scales128 = _mm_set_epi32( - ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), - ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), - (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), - (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); - scales128 = _mm_sub_epi8(scales128, m32); - const __m256i all_scales = _mm256_cvtepi8_epi16(scales128); - const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); - const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); - const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; + const uint32_t * restrict sc = (const uint32_t *)x[i].scales; + ud = (sc[0] >> 0) & 0x0f0f0f0f; + um = (sc[0] >> 4) & 0x0f0f0f0f; - // high bit - const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].hmask); + int32_t smin = mb[0] * y[i].bsums[0] + mb[1] * y[i].bsums[1] + mb[2] * y[i].bsums[2] + mb[3] * y[i].bsums[3]; + summs += dmin * smin; - // integer accumulator - __m256i sumi = _mm256_setzero_si256(); + const __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); + const __m256i q2_0 = _mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q2bits, 2), q2bits), m3); + const __m256i q2_1 = _mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q2bits, 6), _mm_srli_epi16(q2bits, 4)), m3); - int bit = 0; - int is = 0; + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - for (int j = 0; j < QK_K/128; ++j) { - // load low 2 bits - const __m256i q3bits = _mm256_loadu_si256((const __m256i*)q3); q3 += 32; + const __m256i p0 = _mm256_maddubs_epi16(q2_0, q8_0); + const __m256i p1 = _mm256_maddubs_epi16(q2_1, q8_1); - // prepare low and high bits - const __m256i q3l_0 = _mm256_and_si256(q3bits, m3); - const __m256i q3h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; + const __m256i p_0 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p0, 0)); + const __m256i p_1 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p0, 1)); + const __m256i p_2 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p1, 0)); + const __m256i p_3 = _mm256_cvtepi16_epi32(_mm256_extracti128_si256(p1, 1)); - const __m256i q3l_1 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 2), m3); - const __m256i q3h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; + acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[0]), _mm256_cvtepi32_ps(p_0), acc); + acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[1]), _mm256_cvtepi32_ps(p_1), acc); + acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[2]), _mm256_cvtepi32_ps(p_2), acc); + acc = _mm256_fmadd_ps(_mm256_set1_ps(d * db[3]), _mm256_cvtepi32_ps(p_3), acc); + } - const __m256i q3l_2 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 4), m3); - const __m256i q3h_2 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; + *s = hsum_float_8(acc) + summs; - const __m256i q3l_3 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 6), m3); - const __m256i q3h_3 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); - ++bit; +#elif defined __AVX__ - // load Q8 quants - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m128i m3 = _mm_set1_epi8(3); - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - __m256i q8s_0 = _mm256_maddubs_epi16(q3h_0, q8_0); - __m256i q8s_1 = _mm256_maddubs_epi16(q3h_1, q8_1); - __m256i q8s_2 = _mm256_maddubs_epi16(q3h_2, q8_2); - __m256i q8s_3 = _mm256_maddubs_epi16(q3h_3, q8_3); + __m256 acc = _mm256_setzero_ps(); - __m256i p16_0 = _mm256_maddubs_epi16(q3l_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q3l_1, q8_1); - __m256i p16_2 = _mm256_maddubs_epi16(q3l_2, q8_2); - __m256i p16_3 = _mm256_maddubs_epi16(q3l_3, q8_3); + uint32_t ud, um; + const uint8_t * restrict db = (const uint8_t *)&ud; + const uint8_t * restrict mb = (const uint8_t *)&um; - p16_0 = _mm256_sub_epi16(p16_0, q8s_0); - p16_1 = _mm256_sub_epi16(p16_1, q8s_1); - p16_2 = _mm256_sub_epi16(p16_2, q8s_2); - p16_3 = _mm256_sub_epi16(p16_3, q8s_3); + float summs = 0; - // multiply with scales - p16_0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 0)), p16_0); - p16_1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 1)), p16_1); - p16_2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 2)), p16_2); - p16_3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 3)), p16_3); + // TODO: optimize this - // accumulate - p16_0 = _mm256_add_epi32(p16_0, p16_1); - p16_2 = _mm256_add_epi32(p16_2, p16_3); - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_2)); + for (int i = 0; i < nb; ++i) { - } + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - // multiply with block scale and accumulate - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + const uint8_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; - } + const uint32_t * restrict sc = (const uint32_t *)x[i].scales; + ud = (sc[0] >> 0) & 0x0f0f0f0f; + um = (sc[0] >> 4) & 0x0f0f0f0f; - *s = hsum_float_8(acc); + int32_t smin = mb[0] * y[i].bsums[0] + mb[1] * y[i].bsums[1] + mb[2] * y[i].bsums[2] + mb[3] * y[i].bsums[3]; + summs += dmin * smin; -#elif defined __AVX__ + const __m128i q2bits = _mm_loadu_si128((const __m128i*)q2); + const __m128i q2_0 = _mm_and_si128(q2bits, m3); + const __m128i q2_1 = _mm_and_si128(_mm_srli_epi16(q2bits, 2), m3); + const __m128i q2_2 = _mm_and_si128(_mm_srli_epi16(q2bits, 4), m3); + const __m128i q2_3 = _mm_and_si128(_mm_srli_epi16(q2bits, 6), m3); - const __m128i m3 = _mm_set1_epi8(3); - const __m128i mone = _mm_set1_epi8(1); - const __m128i m32 = _mm_set1_epi8(32); - const __m128i m2 = _mm_set1_epi8(2); + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - __m256 acc = _mm256_setzero_ps(); + const __m128i p0 = _mm_maddubs_epi16(q2_0, _mm256_extractf128_si256(q8_0, 0)); + const __m128i p1 = _mm_maddubs_epi16(q2_1, _mm256_extractf128_si256(q8_0, 1)); + const __m128i p2 = _mm_maddubs_epi16(q2_2, _mm256_extractf128_si256(q8_1, 0)); + const __m128i p3 = _mm_maddubs_epi16(q2_3, _mm256_extractf128_si256(q8_1, 1)); - const uint32_t *aux; + const __m256i p_0 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p0, p0)), _mm_cvtepi16_epi32(p0)); + const __m256i p_1 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p1, p1)), _mm_cvtepi16_epi32(p1)); + const __m256i p_2 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p2, p2)), _mm_cvtepi16_epi32(p2)); + const __m256i p_3 = MM256_SET_M128I(_mm_cvtepi16_epi32(_mm_unpackhi_epi64(p3, p3)), _mm_cvtepi16_epi32(p3)); + + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[0]), _mm256_cvtepi32_ps(p_0)), acc); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[1]), _mm256_cvtepi32_ps(p_1)), acc); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[2]), _mm256_cvtepi32_ps(p_2)), acc); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d * db[3]), _mm256_cvtepi32_ps(p_3)), acc); + } + + *s = hsum_float_8(acc) + summs; + +#elif defined __riscv_v_intrinsic + + uint32_t aux32[2]; + const uint8_t * scales = (const uint8_t *)aux32; + + float sumf = 0; for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict q2 = x[i].qs; const int8_t * restrict q8 = y[i].qs; + const uint32_t * restrict sc = (const uint32_t *)x[i].scales; - // Set up scales - aux = (const uint32_t *)x[i].scales; - __m128i scales128 = _mm_set_epi32( - ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), - ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), - (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), - (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); - scales128 = _mm_sub_epi8(scales128, m32); - const __m128i scales_0 = _mm_cvtepi8_epi16(scales128); - const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales128, scales128)); - const __m128i scales[2] = { scales_0, scales_1 }; + aux32[0] = sc[0] & 0x0f0f0f0f; + aux32[1] = (sc[0] >> 4) & 0x0f0f0f0f; - // high bit *128*2 from block_q3_K.hmask[QK_K/8] - const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].hmask[0]); - const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].hmask[16]); + sumf += dmin * (scales[4] * y[i].bsums[0] + scales[5] * y[i].bsums[1] + scales[6] * y[i].bsums[2] + scales[7] * y[i].bsums[3]); - // integer accumulator - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); + int isum1 = 0; + int isum2 = 0; - for (int j = 0; j < QK_K/128; ++j) { - // load low 2 bits *64*2 from block_q3_K.qs[QK_K/4] - const __m128i q3bits_0 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; - const __m128i q3bits_1 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; + size_t vl = 16; - // prepare low and high bits - const int bit = j << 2; + vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); - const __m128i q3l_0 = _mm_and_si128(q3bits_0, m3); - const __m128i q3l_1 = _mm_and_si128(q3bits_1, m3); - const __m128i q3h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit)), bit), 2); - const __m128i q3h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit)), bit), 2); + // load Q2 + vuint8mf2_t q2_x = __riscv_vle8_v_u8mf2(q2, vl); - const __m128i q3l_2 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 2), m3); - const __m128i q3l_3 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 2), m3); - const __m128i q3h_2 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+1)), bit+1), 2); - const __m128i q3h_3 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+1)), bit+1), 2); + vint8mf2_t q2_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(q2_x, 0x03, vl)); + vint8mf2_t q2_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q2_x, 0x2, vl), 0x03 , vl)); + vint8mf2_t q2_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q2_x, 0x4, vl), 0x03 , vl)); + vint8mf2_t q2_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q2_x, 0x6, vl), 0x03 , vl)); - const __m128i q3l_4 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 4), m3); - const __m128i q3l_5 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 4), m3); - const __m128i q3h_4 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+2)), bit+2), 2); - const __m128i q3h_5 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+2)), bit+2), 2); + // load Q8, and take product with Q2 + vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q2_0, __riscv_vle8_v_i8mf2(q8, vl), vl); + vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q2_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); + vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q2_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); + vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q2_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); - const __m128i q3l_6 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 6), m3); - const __m128i q3l_7 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 6), m3); - const __m128i q3h_6 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+3)), bit+3), 2); - const __m128i q3h_7 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+3)), bit+3), 2); + vint16m1_t vs_0 = __riscv_vredsum_vs_i16m1_i16m1(p0, vzero, vl); + vint16m1_t vs_1 = __riscv_vredsum_vs_i16m1_i16m1(p1, vzero, vl); + vint16m1_t vs_2 = __riscv_vredsum_vs_i16m1_i16m1(p2, vzero, vl); + vint16m1_t vs_3 = __riscv_vredsum_vs_i16m1_i16m1(p3, vzero, vl); - // load Q8 quants from block_q8_K.qs[QK_K] - const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + isum1 += __riscv_vmv_x_s_i16m1_i16(vs_0) * scales[0]; + isum2 += __riscv_vmv_x_s_i16m1_i16(vs_1) * scales[1]; + isum1 += __riscv_vmv_x_s_i16m1_i16(vs_2) * scales[2]; + isum2 += __riscv_vmv_x_s_i16m1_i16(vs_3) * scales[3]; - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - __m128i q8s_0 = _mm_maddubs_epi16(q3h_0, q8_0); - __m128i q8s_1 = _mm_maddubs_epi16(q3h_1, q8_1); - __m128i q8s_2 = _mm_maddubs_epi16(q3h_2, q8_2); - __m128i q8s_3 = _mm_maddubs_epi16(q3h_3, q8_3); - __m128i q8s_4 = _mm_maddubs_epi16(q3h_4, q8_4); - __m128i q8s_5 = _mm_maddubs_epi16(q3h_5, q8_5); - __m128i q8s_6 = _mm_maddubs_epi16(q3h_6, q8_6); - __m128i q8s_7 = _mm_maddubs_epi16(q3h_7, q8_7); + sumf += d * (isum1 + isum2); - __m128i p16_0 = _mm_maddubs_epi16(q3l_0, q8_0); - __m128i p16_1 = _mm_maddubs_epi16(q3l_1, q8_1); - __m128i p16_2 = _mm_maddubs_epi16(q3l_2, q8_2); - __m128i p16_3 = _mm_maddubs_epi16(q3l_3, q8_3); - __m128i p16_4 = _mm_maddubs_epi16(q3l_4, q8_4); - __m128i p16_5 = _mm_maddubs_epi16(q3l_5, q8_5); - __m128i p16_6 = _mm_maddubs_epi16(q3l_6, q8_6); - __m128i p16_7 = _mm_maddubs_epi16(q3l_7, q8_7); + } - p16_0 = _mm_sub_epi16(p16_0, q8s_0); - p16_1 = _mm_sub_epi16(p16_1, q8s_1); - p16_2 = _mm_sub_epi16(p16_2, q8s_2); - p16_3 = _mm_sub_epi16(p16_3, q8s_3); - p16_4 = _mm_sub_epi16(p16_4, q8s_4); - p16_5 = _mm_sub_epi16(p16_5, q8s_5); - p16_6 = _mm_sub_epi16(p16_6, q8s_6); - p16_7 = _mm_sub_epi16(p16_7, q8s_7); + *s = sumf; - // multiply with scales - __m128i shuffle = _mm_set1_epi16(0x0100); - p16_0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_0); - shuffle = _mm_add_epi16(shuffle, m2); - p16_1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_1); - shuffle = _mm_add_epi16(shuffle, m2); - p16_2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_2); - shuffle = _mm_add_epi16(shuffle, m2); - p16_3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_3); - shuffle = _mm_add_epi16(shuffle, m2); - p16_4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_4); - shuffle = _mm_add_epi16(shuffle, m2); - p16_5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_5); - shuffle = _mm_add_epi16(shuffle, m2); - p16_6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_6); - shuffle = _mm_add_epi16(shuffle, m2); - p16_7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_7); +#else - // accumulate - p16_0 = _mm_add_epi32(p16_0, p16_1); - p16_2 = _mm_add_epi32(p16_2, p16_3); - p16_4 = _mm_add_epi32(p16_4, p16_5); - p16_6 = _mm_add_epi32(p16_6, p16_7); - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_4, p16_6)); + float sumf = 0; + + int isum[QK_K/16]; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * q2 = x[i].qs; + const int8_t * q8 = y[i].qs; + const uint8_t * sc = x[i].scales; + int summs = 0; + for (int j = 0; j < QK_K/16; ++j) { + summs += y[i].bsums[j] * (sc[j] >> 4); } - // multiply with block scale and accumulate - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi)), acc); + const float dall = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + memset(isum, 0, (QK_K/16)*sizeof(int)); + for (int l = 0; l < 16; ++l) { + isum[0] += q8[l+ 0] * ((q2[l] >> 0) & 3); + isum[1] += q8[l+16] * ((q2[l] >> 2) & 3); + isum[2] += q8[l+32] * ((q2[l] >> 4) & 3); + isum[3] += q8[l+48] * ((q2[l] >> 6) & 3); + } + for (int l = 0; l < QK_K/16; ++l) { + isum[l] *= (sc[l] & 0xF); + } + sumf += dall * (isum[0] + isum[1] + isum[2] + isum[3]) - dmin * summs; } + *s = sumf; +#endif +} +#endif - *s = hsum_float_8(acc); +#if QK_K == 256 +void ggml_vec_dot_q3_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); -#elif defined __riscv_v_intrinsic + const uint32_t kmask1 = 0x03030303; + const uint32_t kmask2 = 0x0f0f0f0f; + + const block_q3_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#ifdef __ARM_NEON uint32_t aux[3]; uint32_t utmp[4]; - float sumf = 0; + const uint8x16_t m3b = vdupq_n_u8(0x3); + const int32x4_t vzero = vdupq_n_s32(0); + + const uint8x16_t m0 = vdupq_n_u8(1); + const uint8x16_t m1 = vshlq_n_u8(m0, 1); + const uint8x16_t m2 = vshlq_n_u8(m0, 2); + const uint8x16_t m3 = vshlq_n_u8(m0, 3); + const int8_t m32 = 32; + + ggml_int8x16x4_t q3bytes; + + float sum = 0; + for (int i = 0; i < nb; ++i) { + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const uint8_t * restrict q3 = x[i].qs; const uint8_t * restrict qh = x[i].hmask; - const int8_t * restrict q8 = y[i].qs; + const int8_t * restrict q8 = y[i].qs; + ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); + + ggml_uint8x16x4_t q3h; + + int32_t isum = 0; + + // Set up scales memcpy(aux, x[i].scales, 12); utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); @@ -4605,90 +5728,409 @@ void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restri utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); int8_t * scale = (int8_t *)utmp; - for (int j = 0; j < 16; ++j) scale[j] -= 32; - - - size_t vl = 32; - uint8_t m = 1; + for (int j = 0; j < 16; ++j) scale[j] -= m32; - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vuint8m1_t vqh = __riscv_vle8_v_u8m1(qh, vl); + for (int j = 0; j < QK_K/128; ++j) { - int sum_t = 0; + const ggml_uint8x16x2_t q3bits = ggml_vld1q_u8_x2(q3); q3 += 32; + const ggml_int8x16x4_t q8bytes_1 = ggml_vld1q_s8_x4(q8); q8 += 64; + const ggml_int8x16x4_t q8bytes_2 = ggml_vld1q_s8_x4(q8); q8 += 64; - for (int j = 0; j < QK_K; j += 128) { + q3h.val[0] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[0]), 2); + q3h.val[1] = vshlq_n_u8(vbicq_u8(m0, qhbits.val[1]), 2); + q3h.val[2] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[0]), 1); + q3h.val[3] = vshlq_n_u8(vbicq_u8(m1, qhbits.val[1]), 1); - vl = 32; + q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[0], m3b)), vreinterpretq_s8_u8(q3h.val[0])); + q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q3bits.val[1], m3b)), vreinterpretq_s8_u8(q3h.val[1])); + q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 2), m3b)), vreinterpretq_s8_u8(q3h.val[2])); + q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 2), m3b)), vreinterpretq_s8_u8(q3h.val[3])); - // load Q3 - vuint8m1_t q3_x = __riscv_vle8_v_u8m1(q3, vl); + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes_1.val[0])) * scale[0]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes_1.val[1])) * scale[1]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes_1.val[2])) * scale[2]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes_1.val[3])) * scale[3]; - vint8m1_t q3_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q3_x, 0x03, vl)); - vint8m1_t q3_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x2, vl), 0x03 , vl)); - vint8m1_t q3_2 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x4, vl), 0x03 , vl)); - vint8m1_t q3_3 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x6, vl), 0x03 , vl)); + scale += 4; - // compute mask for subtraction - vuint8m1_t qh_m0 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_0 = __riscv_vmseq_vx_u8m1_b8(qh_m0, 0, vl); - vint8m1_t q3_m0 = __riscv_vsub_vx_i8m1_m(vmask_0, q3_0, 0x4, vl); - m <<= 1; + q3h.val[0] = vbicq_u8(m2, qhbits.val[0]); + q3h.val[1] = vbicq_u8(m2, qhbits.val[1]); + q3h.val[2] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[0]), 1); + q3h.val[3] = vshrq_n_u8(vbicq_u8(m3, qhbits.val[1]), 1); - vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_1 = __riscv_vmseq_vx_u8m1_b8(qh_m1, 0, vl); - vint8m1_t q3_m1 = __riscv_vsub_vx_i8m1_m(vmask_1, q3_1, 0x4, vl); - m <<= 1; + q3bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 4), m3b)), vreinterpretq_s8_u8(q3h.val[0])); + q3bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 4), m3b)), vreinterpretq_s8_u8(q3h.val[1])); + q3bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[0], 6), m3b)), vreinterpretq_s8_u8(q3h.val[2])); + q3bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(vshrq_n_u8(q3bits.val[1], 6), m3b)), vreinterpretq_s8_u8(q3h.val[3])); - vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_2 = __riscv_vmseq_vx_u8m1_b8(qh_m2, 0, vl); - vint8m1_t q3_m2 = __riscv_vsub_vx_i8m1_m(vmask_2, q3_2, 0x4, vl); - m <<= 1; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes_2.val[0])) * scale[0]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes_2.val[1])) * scale[1]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes_2.val[2])) * scale[2]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes_2.val[3])) * scale[3]; - vuint8m1_t qh_m3 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_3 = __riscv_vmseq_vx_u8m1_b8(qh_m3, 0, vl); - vint8m1_t q3_m3 = __riscv_vsub_vx_i8m1_m(vmask_3, q3_3, 0x4, vl); - m <<= 1; + scale += 4; - // load Q8 and take product with Q3 - vint16m2_t a0 = __riscv_vwmul_vv_i16m2(q3_m0, __riscv_vle8_v_i8m1(q8, vl), vl); - vint16m2_t a1 = __riscv_vwmul_vv_i16m2(q3_m1, __riscv_vle8_v_i8m1(q8+32, vl), vl); - vint16m2_t a2 = __riscv_vwmul_vv_i16m2(q3_m2, __riscv_vle8_v_i8m1(q8+64, vl), vl); - vint16m2_t a3 = __riscv_vwmul_vv_i16m2(q3_m3, __riscv_vle8_v_i8m1(q8+96, vl), vl); + if (j == 0) { + qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 4); + qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 4); + } - vl = 16; + } + sum += d * isum; - // retreive lane to multiply with scale - vint32m2_t aux0_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 0), (scale[0]), vl); - vint32m2_t aux0_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 1), (scale[1]), vl); - vint32m2_t aux1_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 0), (scale[2]), vl); - vint32m2_t aux1_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 1), (scale[3]), vl); - vint32m2_t aux2_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 0), (scale[4]), vl); - vint32m2_t aux2_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 1), (scale[5]), vl); - vint32m2_t aux3_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 0), (scale[6]), vl); - vint32m2_t aux3_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 1), (scale[7]), vl); + } - vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux0_0, aux0_1, vl), vzero, vl); - vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux1_0, aux1_1, vl), isum0, vl); - vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux2_0, aux2_1, vl), isum1, vl); - vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux3_0, aux3_1, vl), isum2, vl); + *s = sum; - sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); +#elif defined __AVX2__ - q3 += 32; q8 += 128; scale += 8; + const __m256i m3 = _mm256_set1_epi8(3); + const __m256i mone = _mm256_set1_epi8(1); + const __m128i m32 = _mm_set1_epi8(32); - } + __m256 acc = _mm256_setzero_ps(); - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + uint32_t aux[3]; - sumf += d*sum_t; + for (int i = 0; i < nb; ++i) { - } + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - *s = sumf; + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; -#else - // scalar version - // This function is written like this so the compiler can manage to vectorize most of it + // Set up scales + memcpy(aux, x[i].scales, 12); + __m128i scales128 = _mm_set_epi32( + ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), + ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), + (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), + (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); + scales128 = _mm_sub_epi8(scales128, m32); + const __m256i all_scales = _mm256_cvtepi8_epi16(scales128); + const __m128i l_scales = _mm256_extracti128_si256(all_scales, 0); + const __m128i h_scales = _mm256_extracti128_si256(all_scales, 1); + const __m256i scales[2] = {MM256_SET_M128I(l_scales, l_scales), MM256_SET_M128I(h_scales, h_scales)}; + + // high bit + const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].hmask); + + // integer accumulator + __m256i sumi = _mm256_setzero_si256(); + + int bit = 0; + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + // load low 2 bits + const __m256i q3bits = _mm256_loadu_si256((const __m256i*)q3); q3 += 32; + + // prepare low and high bits + const __m256i q3l_0 = _mm256_and_si256(q3bits, m3); + const __m256i q3h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + const __m256i q3l_1 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 2), m3); + const __m256i q3h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + const __m256i q3l_2 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 4), m3); + const __m256i q3h_2 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + const __m256i q3l_3 = _mm256_and_si256(_mm256_srli_epi16(q3bits, 6), m3); + const __m256i q3h_3 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_andnot_si256(hbits, _mm256_slli_epi16(mone, bit)), bit), 2); + ++bit; + + // load Q8 quants + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + __m256i q8s_0 = _mm256_maddubs_epi16(q3h_0, q8_0); + __m256i q8s_1 = _mm256_maddubs_epi16(q3h_1, q8_1); + __m256i q8s_2 = _mm256_maddubs_epi16(q3h_2, q8_2); + __m256i q8s_3 = _mm256_maddubs_epi16(q3h_3, q8_3); + + __m256i p16_0 = _mm256_maddubs_epi16(q3l_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q3l_1, q8_1); + __m256i p16_2 = _mm256_maddubs_epi16(q3l_2, q8_2); + __m256i p16_3 = _mm256_maddubs_epi16(q3l_3, q8_3); + + p16_0 = _mm256_sub_epi16(p16_0, q8s_0); + p16_1 = _mm256_sub_epi16(p16_1, q8s_1); + p16_2 = _mm256_sub_epi16(p16_2, q8s_2); + p16_3 = _mm256_sub_epi16(p16_3, q8s_3); + + // multiply with scales + p16_0 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 0)), p16_0); + p16_1 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 1)), p16_1); + p16_2 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 2)), p16_2); + p16_3 = _mm256_madd_epi16(_mm256_shuffle_epi8(scales[j], get_scale_shuffle_q3k(is + 3)), p16_3); + + // accumulate + p16_0 = _mm256_add_epi32(p16_0, p16_1); + p16_2 = _mm256_add_epi32(p16_2, p16_3); + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_2)); + + } + + // multiply with block scale and accumulate + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __AVX__ + + const __m128i m3 = _mm_set1_epi8(3); + const __m128i mone = _mm_set1_epi8(1); + const __m128i m32 = _mm_set1_epi8(32); + const __m128i m2 = _mm_set1_epi8(2); + + __m256 acc = _mm256_setzero_ps(); + + const uint32_t *aux; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + // Set up scales + aux = (const uint32_t *)x[i].scales; + __m128i scales128 = _mm_set_epi32( + ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4), + ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4), + (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4), + (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4)); + scales128 = _mm_sub_epi8(scales128, m32); + const __m128i scales_0 = _mm_cvtepi8_epi16(scales128); + const __m128i scales_1 = _mm_cvtepi8_epi16(_mm_unpackhi_epi64(scales128, scales128)); + const __m128i scales[2] = { scales_0, scales_1 }; + + // high bit *128*2 from block_q3_K.hmask[QK_K/8] + const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].hmask[0]); + const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].hmask[16]); + + // integer accumulator + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + for (int j = 0; j < QK_K/128; ++j) { + // load low 2 bits *64*2 from block_q3_K.qs[QK_K/4] + const __m128i q3bits_0 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; + const __m128i q3bits_1 = _mm_loadu_si128((const __m128i*)q3); q3 += 16; + + // prepare low and high bits + const int bit = j << 2; + + const __m128i q3l_0 = _mm_and_si128(q3bits_0, m3); + const __m128i q3l_1 = _mm_and_si128(q3bits_1, m3); + const __m128i q3h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit)), bit), 2); + const __m128i q3h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit)), bit), 2); + + const __m128i q3l_2 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 2), m3); + const __m128i q3l_3 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 2), m3); + const __m128i q3h_2 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+1)), bit+1), 2); + const __m128i q3h_3 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+1)), bit+1), 2); + + const __m128i q3l_4 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 4), m3); + const __m128i q3l_5 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 4), m3); + const __m128i q3h_4 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+2)), bit+2), 2); + const __m128i q3h_5 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+2)), bit+2), 2); + + const __m128i q3l_6 = _mm_and_si128(_mm_srli_epi16(q3bits_0, 6), m3); + const __m128i q3l_7 = _mm_and_si128(_mm_srli_epi16(q3bits_1, 6), m3); + const __m128i q3h_6 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_0, _mm_slli_epi16(mone, bit+3)), bit+3), 2); + const __m128i q3h_7 = _mm_slli_epi16(_mm_srli_epi16(_mm_andnot_si128(hbits_1, _mm_slli_epi16(mone, bit+3)), bit+3), 2); + + // load Q8 quants from block_q8_K.qs[QK_K] + const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + __m128i q8s_0 = _mm_maddubs_epi16(q3h_0, q8_0); + __m128i q8s_1 = _mm_maddubs_epi16(q3h_1, q8_1); + __m128i q8s_2 = _mm_maddubs_epi16(q3h_2, q8_2); + __m128i q8s_3 = _mm_maddubs_epi16(q3h_3, q8_3); + __m128i q8s_4 = _mm_maddubs_epi16(q3h_4, q8_4); + __m128i q8s_5 = _mm_maddubs_epi16(q3h_5, q8_5); + __m128i q8s_6 = _mm_maddubs_epi16(q3h_6, q8_6); + __m128i q8s_7 = _mm_maddubs_epi16(q3h_7, q8_7); + + __m128i p16_0 = _mm_maddubs_epi16(q3l_0, q8_0); + __m128i p16_1 = _mm_maddubs_epi16(q3l_1, q8_1); + __m128i p16_2 = _mm_maddubs_epi16(q3l_2, q8_2); + __m128i p16_3 = _mm_maddubs_epi16(q3l_3, q8_3); + __m128i p16_4 = _mm_maddubs_epi16(q3l_4, q8_4); + __m128i p16_5 = _mm_maddubs_epi16(q3l_5, q8_5); + __m128i p16_6 = _mm_maddubs_epi16(q3l_6, q8_6); + __m128i p16_7 = _mm_maddubs_epi16(q3l_7, q8_7); + + p16_0 = _mm_sub_epi16(p16_0, q8s_0); + p16_1 = _mm_sub_epi16(p16_1, q8s_1); + p16_2 = _mm_sub_epi16(p16_2, q8s_2); + p16_3 = _mm_sub_epi16(p16_3, q8s_3); + p16_4 = _mm_sub_epi16(p16_4, q8s_4); + p16_5 = _mm_sub_epi16(p16_5, q8s_5); + p16_6 = _mm_sub_epi16(p16_6, q8s_6); + p16_7 = _mm_sub_epi16(p16_7, q8s_7); + + // multiply with scales + __m128i shuffle = _mm_set1_epi16(0x0100); + p16_0 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_0); + shuffle = _mm_add_epi16(shuffle, m2); + p16_1 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_1); + shuffle = _mm_add_epi16(shuffle, m2); + p16_2 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_2); + shuffle = _mm_add_epi16(shuffle, m2); + p16_3 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_3); + shuffle = _mm_add_epi16(shuffle, m2); + p16_4 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_4); + shuffle = _mm_add_epi16(shuffle, m2); + p16_5 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_5); + shuffle = _mm_add_epi16(shuffle, m2); + p16_6 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_6); + shuffle = _mm_add_epi16(shuffle, m2); + p16_7 = _mm_madd_epi16(_mm_shuffle_epi8(scales[j], shuffle), p16_7); + + // accumulate + p16_0 = _mm_add_epi32(p16_0, p16_1); + p16_2 = _mm_add_epi32(p16_2, p16_3); + p16_4 = _mm_add_epi32(p16_4, p16_5); + p16_6 = _mm_add_epi32(p16_6, p16_7); + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_4, p16_6)); + + } + + // multiply with block scale and accumulate + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi)), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __riscv_v_intrinsic + + uint32_t aux[3]; + uint32_t utmp[4]; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict qh = x[i].hmask; + const int8_t * restrict q8 = y[i].qs; + + memcpy(aux, x[i].scales, 12); + utmp[3] = ((aux[1] >> 4) & kmask2) | (((aux[2] >> 6) & kmask1) << 4); + utmp[2] = ((aux[0] >> 4) & kmask2) | (((aux[2] >> 4) & kmask1) << 4); + utmp[1] = (aux[1] & kmask2) | (((aux[2] >> 2) & kmask1) << 4); + utmp[0] = (aux[0] & kmask2) | (((aux[2] >> 0) & kmask1) << 4); + + int8_t * scale = (int8_t *)utmp; + for (int j = 0; j < 16; ++j) scale[j] -= 32; + + + size_t vl = 32; + uint8_t m = 1; + + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + vuint8m1_t vqh = __riscv_vle8_v_u8m1(qh, vl); + + int sum_t = 0; + + for (int j = 0; j < QK_K; j += 128) { + + vl = 32; + + // load Q3 + vuint8m1_t q3_x = __riscv_vle8_v_u8m1(q3, vl); + + vint8m1_t q3_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q3_x, 0x03, vl)); + vint8m1_t q3_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x2, vl), 0x03 , vl)); + vint8m1_t q3_2 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x4, vl), 0x03 , vl)); + vint8m1_t q3_3 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(q3_x, 0x6, vl), 0x03 , vl)); + + // compute mask for subtraction + vuint8m1_t qh_m0 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_0 = __riscv_vmseq_vx_u8m1_b8(qh_m0, 0, vl); + vint8m1_t q3_m0 = __riscv_vsub_vx_i8m1_m(vmask_0, q3_0, 0x4, vl); + m <<= 1; + + vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_1 = __riscv_vmseq_vx_u8m1_b8(qh_m1, 0, vl); + vint8m1_t q3_m1 = __riscv_vsub_vx_i8m1_m(vmask_1, q3_1, 0x4, vl); + m <<= 1; + + vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_2 = __riscv_vmseq_vx_u8m1_b8(qh_m2, 0, vl); + vint8m1_t q3_m2 = __riscv_vsub_vx_i8m1_m(vmask_2, q3_2, 0x4, vl); + m <<= 1; + + vuint8m1_t qh_m3 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_3 = __riscv_vmseq_vx_u8m1_b8(qh_m3, 0, vl); + vint8m1_t q3_m3 = __riscv_vsub_vx_i8m1_m(vmask_3, q3_3, 0x4, vl); + m <<= 1; + + // load Q8 and take product with Q3 + vint16m2_t a0 = __riscv_vwmul_vv_i16m2(q3_m0, __riscv_vle8_v_i8m1(q8, vl), vl); + vint16m2_t a1 = __riscv_vwmul_vv_i16m2(q3_m1, __riscv_vle8_v_i8m1(q8+32, vl), vl); + vint16m2_t a2 = __riscv_vwmul_vv_i16m2(q3_m2, __riscv_vle8_v_i8m1(q8+64, vl), vl); + vint16m2_t a3 = __riscv_vwmul_vv_i16m2(q3_m3, __riscv_vle8_v_i8m1(q8+96, vl), vl); + + vl = 16; + + // retrieve lane to multiply with scale + vint32m2_t aux0_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 0), (scale[0]), vl); + vint32m2_t aux0_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a0, 1), (scale[1]), vl); + vint32m2_t aux1_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 0), (scale[2]), vl); + vint32m2_t aux1_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a1, 1), (scale[3]), vl); + vint32m2_t aux2_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 0), (scale[4]), vl); + vint32m2_t aux2_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a2, 1), (scale[5]), vl); + vint32m2_t aux3_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 0), (scale[6]), vl); + vint32m2_t aux3_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(a3, 1), (scale[7]), vl); + + vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux0_0, aux0_1, vl), vzero, vl); + vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux1_0, aux1_1, vl), isum0, vl); + vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux2_0, aux2_1, vl), isum1, vl); + vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(aux3_0, aux3_1, vl), isum2, vl); + + sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); + + q3 += 32; q8 += 128; scale += 8; + + } + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + + sumf += d*sum_t; + + } + + *s = sumf; + +#else + // scalar version + // This function is written like this so the compiler can manage to vectorize most of it // Using -Ofast, GCC and clang manage to produce code that is within a factor of 2 or so from the // manually vectorized version above. Every other version I tried would run at least 4 times slower. // The ideal situation would be if we could just write the code once, and the compiler would @@ -4701,202 +6143,1541 @@ void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restri int32_t aux32[8]; memset(sums, 0, 8*sizeof(float)); - uint32_t auxs[4]; - const int8_t * scales = (const int8_t*)auxs; - + uint32_t auxs[4]; + const int8_t * scales = (const int8_t*)auxs; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict hm = x[i].hmask; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + uint8_t m = 1; + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) a[l] = q3[l] & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 2) & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 4) & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 6) & 3; + for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + a += 32; m <<= 1; + q3 += 32; + } + a = aux8; + + memcpy(auxs, x[i].scales, 12); + uint32_t tmp = auxs[2]; + auxs[2] = ((auxs[0] >> 4) & kmask2) | (((tmp >> 4) & kmask1) << 4); + auxs[3] = ((auxs[1] >> 4) & kmask2) | (((tmp >> 6) & kmask1) << 4); + auxs[0] = (auxs[0] & kmask2) | (((tmp >> 0) & kmask1) << 4); + auxs[1] = (auxs[1] & kmask2) | (((tmp >> 2) & kmask1) << 4); + for (int j = 0; j < QK_K/16; ++j) { + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; + q8 += 8; a += 8; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; + +#endif + +} + +#else + +void ggml_vec_dot_q3_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q3_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#ifdef __ARM_NEON + const int32x4_t vzero = vdupq_n_s32(0); + + const uint8x16_t m3b = vdupq_n_u8(0x3); + const uint8x16_t mh = vdupq_n_u8(4); + + ggml_int8x16x4_t q3bytes; + + uint16_t aux16[2]; + int8_t * scales = (int8_t *)aux16; + + float sum = 0; + + for (int i = 0; i < nb; ++i) { + + ggml_uint8x16x4_t q3h; + + const uint8x8_t hbits = vld1_u8(x[i].hmask); + const uint8x16_t q3bits = vld1q_u8(x[i].qs); + const ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(y[i].qs); + + const uint16_t a = *(const uint16_t *)x[i].scales; + aux16[0] = a & 0x0f0f; + aux16[1] = (a >> 4) & 0x0f0f; + + for (int j = 0; j < 4; ++j) scales[j] -= 8; + + int32_t isum = -4*(scales[0] * y[i].bsums[0] + scales[2] * y[i].bsums[1] + scales[1] * y[i].bsums[2] + scales[3] * y[i].bsums[3]); + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8x16_t htmp = vcombine_u8(hbits, vshr_n_u8(hbits, 1)); + q3h.val[0] = vandq_u8(mh, vshlq_n_u8(htmp, 2)); + q3h.val[1] = vandq_u8(mh, htmp); + q3h.val[2] = vandq_u8(mh, vshrq_n_u8(htmp, 2)); + q3h.val[3] = vandq_u8(mh, vshrq_n_u8(htmp, 4)); + + q3bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q3bits, m3b), q3h.val[0])); + q3bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(vshrq_n_u8(q3bits, 2), m3b), q3h.val[1])); + q3bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(vshrq_n_u8(q3bits, 4), m3b), q3h.val[2])); + q3bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q3bits, 6), q3h.val[3])); + + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[0], q8bytes.val[0])) * scales[0]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[1], q8bytes.val[1])) * scales[2]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[2], q8bytes.val[2])) * scales[1]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q3bytes.val[3], q8bytes.val[3])) * scales[3]; + + sum += d * isum; + + } + + *s = sum; + +#elif defined __AVX2__ + + const __m256i m3 = _mm256_set1_epi8(3); + const __m256i m1 = _mm256_set1_epi8(1); + + __m256 acc = _mm256_setzero_ps(); + + uint64_t aux64; + + uint16_t aux16[2]; + const int8_t * aux8 = (const int8_t *)aux16; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const uint16_t a = *(const uint16_t *)x[i].scales; + aux16[0] = a & 0x0f0f; + aux16[1] = (a >> 4) & 0x0f0f; + + const __m256i scale_0 = MM256_SET_M128I(_mm_set1_epi16(aux8[2] - 8), _mm_set1_epi16(aux8[0] - 8)); + const __m256i scale_1 = MM256_SET_M128I(_mm_set1_epi16(aux8[3] - 8), _mm_set1_epi16(aux8[1] - 8)); + + memcpy(&aux64, x[i].hmask, 8); + + const __m128i haux = _mm_set_epi64x(aux64 >> 1, aux64 >> 0); + __m256i q3h_0 = MM256_SET_M128I(_mm_srli_epi16(haux, 2), haux); + __m256i q3h_1 = _mm256_srli_epi16(q3h_0, 4); + q3h_0 = _mm256_slli_epi16(_mm256_andnot_si256(q3h_0, m1), 2); + q3h_1 = _mm256_slli_epi16(_mm256_andnot_si256(q3h_1, m1), 2); + + // load low 2 bits + const __m128i q3bits = _mm_loadu_si128((const __m128i*)q3); + + // prepare low and high bits + const __m256i q3aux = MM256_SET_M128I(_mm_srli_epi16(q3bits, 2), q3bits); + const __m256i q3l_0 = _mm256_and_si256(q3aux, m3); + const __m256i q3l_1 = _mm256_and_si256(_mm256_srli_epi16(q3aux, 4), m3); + + // load Q8 quants + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + const __m256i q8s_0 = _mm256_maddubs_epi16(q3h_0, q8_0); + const __m256i q8s_1 = _mm256_maddubs_epi16(q3h_1, q8_1); + + __m256i p16_0 = _mm256_maddubs_epi16(q3l_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q3l_1, q8_1); + + p16_0 = _mm256_sub_epi16(p16_0, q8s_0); + p16_1 = _mm256_sub_epi16(p16_1, q8s_1); + + // multiply with scales + p16_0 = _mm256_madd_epi16(scale_0, p16_0); + p16_1 = _mm256_madd_epi16(scale_1, p16_1); + + p16_0 = _mm256_add_epi32(p16_0, p16_1); + + // multiply with block scale and accumulate + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(p16_0), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __AVX__ + + const __m128i m3 = _mm_set1_epi8(3); + const __m128i m1 = _mm_set1_epi8(1); + + __m256 acc = _mm256_setzero_ps(); + + uint64_t aux64; + + uint16_t aux16[2]; + const int8_t * aux8 = (const int8_t *)aux16; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const uint16_t a = *(const uint16_t *)x[i].scales; + aux16[0] = a & 0x0f0f; + aux16[1] = (a >> 4) & 0x0f0f; + + const __m128i scale_0 = _mm_set1_epi16(aux8[0] - 8); + const __m128i scale_1 = _mm_set1_epi16(aux8[2] - 8); + const __m128i scale_2 = _mm_set1_epi16(aux8[1] - 8); + const __m128i scale_3 = _mm_set1_epi16(aux8[3] - 8); + + memcpy(&aux64, x[i].hmask, 8); + + __m128i q3h_0 = _mm_set_epi64x(aux64 >> 1, aux64 >> 0); + __m128i q3h_1 = _mm_srli_epi16(q3h_0, 2); + __m128i q3h_2 = _mm_srli_epi16(q3h_0, 4); + __m128i q3h_3 = _mm_srli_epi16(q3h_0, 6); + q3h_0 = _mm_slli_epi16(_mm_andnot_si128(q3h_0, m1), 2); + q3h_1 = _mm_slli_epi16(_mm_andnot_si128(q3h_1, m1), 2); + q3h_2 = _mm_slli_epi16(_mm_andnot_si128(q3h_2, m1), 2); + q3h_3 = _mm_slli_epi16(_mm_andnot_si128(q3h_3, m1), 2); + + // load low 2 bits + const __m128i q3bits = _mm_loadu_si128((const __m128i*)q3); + + // prepare low and high bits + const __m128i q3l_0 = _mm_and_si128(q3bits, m3); + const __m128i q3l_1 = _mm_and_si128(_mm_srli_epi16(q3bits, 2), m3); + const __m128i q3l_2 = _mm_and_si128(_mm_srli_epi16(q3bits, 4), m3); + const __m128i q3l_3 = _mm_and_si128(_mm_srli_epi16(q3bits, 6), m3); + + // load Q8 quants + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + + // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm_maddubs_epi16, + // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, + // and 2 if the high bit was set) + const __m128i q8s_0 = _mm_maddubs_epi16(q3h_0, _mm256_extractf128_si256(q8_0, 0)); + const __m128i q8s_1 = _mm_maddubs_epi16(q3h_1, _mm256_extractf128_si256(q8_0, 1)); + const __m128i q8s_2 = _mm_maddubs_epi16(q3h_2, _mm256_extractf128_si256(q8_1, 0)); + const __m128i q8s_3 = _mm_maddubs_epi16(q3h_3, _mm256_extractf128_si256(q8_1, 1)); + + __m128i p16_0 = _mm_maddubs_epi16(q3l_0, _mm256_extractf128_si256(q8_0, 0)); + __m128i p16_1 = _mm_maddubs_epi16(q3l_1, _mm256_extractf128_si256(q8_0, 1)); + __m128i p16_2 = _mm_maddubs_epi16(q3l_2, _mm256_extractf128_si256(q8_1, 0)); + __m128i p16_3 = _mm_maddubs_epi16(q3l_3, _mm256_extractf128_si256(q8_1, 1)); + + p16_0 = _mm_sub_epi16(p16_0, q8s_0); + p16_1 = _mm_sub_epi16(p16_1, q8s_1); + p16_2 = _mm_sub_epi16(p16_2, q8s_2); + p16_3 = _mm_sub_epi16(p16_3, q8s_3); + + // multiply with scales + p16_0 = _mm_madd_epi16(scale_0, p16_0); + p16_1 = _mm_madd_epi16(scale_1, p16_1); + p16_2 = _mm_madd_epi16(scale_2, p16_2); + p16_3 = _mm_madd_epi16(scale_3, p16_3); + + p16_0 = _mm_add_epi32(p16_0, p16_2); + p16_1 = _mm_add_epi32(p16_1, p16_3); + __m256i p16 = MM256_SET_M128I(p16_1, p16_0); + + // multiply with block scale and accumulate + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(p16)), acc); + + } + + *s = hsum_float_8(acc); + +#elif defined __riscv_v_intrinsic + + uint16_t aux16[2]; + int8_t * scales = (int8_t *)aux16; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q3 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const uint16_t a = *(const uint16_t *)x[i].scales; + aux16[0] = a & 0x0f0f; + aux16[1] = (a >> 4) & 0x0f0f; + + for (int j = 0; j < 4; ++j) scales[j] -= 8; + + int32_t isum = -4*(scales[0] * y[i].bsums[0] + scales[2] * y[i].bsums[1] + scales[1] * y[i].bsums[2] + scales[3] * y[i].bsums[3]); + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + + // load qh + vuint8mf4_t qh_x1 = __riscv_vle8_v_u8mf4(x[i].hmask, 8); + vuint8mf2_t qh_x2 = __riscv_vlmul_ext_v_u8mf4_u8mf2(__riscv_vsrl_vx_u8mf4(qh_x1, 1, 8)); + + size_t vl = 16; + + // extend and combine both qh_x1 and qh_x2 + vuint8mf2_t qh_x = __riscv_vslideup_vx_u8mf2(__riscv_vlmul_ext_v_u8mf4_u8mf2(qh_x1), qh_x2, vl/2, vl); + + vuint8mf2_t qh_0 = __riscv_vand_vx_u8mf2(__riscv_vsll_vx_u8mf2(qh_x, 0x2, vl), 0x4, vl); + vuint8mf2_t qh_1 = __riscv_vand_vx_u8mf2(qh_x, 0x4, vl); + vuint8mf2_t qh_2 = __riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl), 0x4, vl); + vuint8mf2_t qh_3 = __riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(qh_x, 0x4, vl), 0x4, vl); + + // load Q3 + vuint8mf2_t q3_x = __riscv_vle8_v_u8mf2(q3, vl); + + vuint8mf2_t q3h_0 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(q3_x, 0x3, vl), qh_0, vl); + vuint8mf2_t q3h_1 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q3_x, 2, vl), 0x3, vl), qh_1, vl); + vuint8mf2_t q3h_2 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q3_x, 4, vl), 0x3, vl), qh_2, vl); + vuint8mf2_t q3h_3 = __riscv_vor_vv_u8mf2(__riscv_vsrl_vx_u8mf2(q3_x, 0x6, vl), qh_3, vl); + + vint8mf2_t q3_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_0); + vint8mf2_t q3_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_1); + vint8mf2_t q3_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_2); + vint8mf2_t q3_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_3); + + // load Q8 and take product with Q3 + vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q3_0, __riscv_vle8_v_i8mf2(q8, vl), vl); + vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q3_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); + vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q3_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); + vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q3_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); + + vint32m1_t vs_0 = __riscv_vwredsum_vs_i16m1_i32m1(p0, vzero, vl); + vint32m1_t vs_1 = __riscv_vwredsum_vs_i16m1_i32m1(p1, vzero, vl); + vint32m1_t vs_2 = __riscv_vwredsum_vs_i16m1_i32m1(p2, vzero, vl); + vint32m1_t vs_3 = __riscv_vwredsum_vs_i16m1_i32m1(p3, vzero, vl); + + isum += __riscv_vmv_x_s_i32m1_i32(vs_0) * scales[0]; + isum += __riscv_vmv_x_s_i32m1_i32(vs_1) * scales[2]; + isum += __riscv_vmv_x_s_i32m1_i32(vs_2) * scales[1]; + isum += __riscv_vmv_x_s_i32m1_i32(vs_3) * scales[3]; + + sumf += d * isum; + + } + + *s = sumf; + +#else + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + int32_t scales[4]; + memset(sums, 0, 8*sizeof(float)); + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict hm = x[i].hmask; + const int8_t * restrict q8 = y[i].qs; + int8_t * restrict a = aux8; + for (int l = 0; l < 8; ++l) { + a[l+ 0] = (int8_t)((q3[l+0] >> 0) & 3) - (hm[l] & 0x01 ? 0 : 4); + a[l+ 8] = (int8_t)((q3[l+8] >> 0) & 3) - (hm[l] & 0x02 ? 0 : 4); + a[l+16] = (int8_t)((q3[l+0] >> 2) & 3) - (hm[l] & 0x04 ? 0 : 4); + a[l+24] = (int8_t)((q3[l+8] >> 2) & 3) - (hm[l] & 0x08 ? 0 : 4); + a[l+32] = (int8_t)((q3[l+0] >> 4) & 3) - (hm[l] & 0x10 ? 0 : 4); + a[l+40] = (int8_t)((q3[l+8] >> 4) & 3) - (hm[l] & 0x20 ? 0 : 4); + a[l+48] = (int8_t)((q3[l+0] >> 6) & 3) - (hm[l] & 0x40 ? 0 : 4); + a[l+56] = (int8_t)((q3[l+8] >> 6) & 3) - (hm[l] & 0x80 ? 0 : 4); + } + + scales[0] = (x[i].scales[0] & 0xF) - 8; + scales[1] = (x[i].scales[0] >> 4) - 8; + scales[2] = (x[i].scales[1] & 0xF) - 8; + scales[3] = (x[i].scales[1] >> 4) - 8; + + memset(aux32, 0, 8*sizeof(int32_t)); + for (int j = 0; j < QK_K/16; ++j) { + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] += q8[l] * a[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux32[l] += scales[j] * aux16[l]; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; + +#endif + +} +#endif + +#if QK_K == 256 +void ggml_vec_dot_q4_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q4_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + + static const uint32_t kmask1 = 0x3f3f3f3f; + static const uint32_t kmask2 = 0x0f0f0f0f; + static const uint32_t kmask3 = 0x03030303; + + uint32_t utmp[4]; + +#ifdef __ARM_NEON + const uint8x16_t m4b = vdupq_n_u8(0xf); + const int32x4_t mzero = vdupq_n_s32(0); + + ggml_int8x16x2_t q4bytes; + ggml_int8x16x2_t q8bytes; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); + + memcpy(utmp, x[i].scales, 12); + + uint32x2_t mins8 = { 0 }; + mins8 = vset_lane_u32(utmp[1] & kmask1, mins8, 0); + mins8 = vset_lane_u32(((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4), mins8, 1); + + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[0] &= kmask1; + + const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(vreinterpret_u8_u32(mins8))); + const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), + vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); + sumf -= dmin * vaddvq_s32(prod); + + const uint8_t * scales = (const uint8_t *)utmp; + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + int32_t sumi1 = 0; + int32_t sumi2 = 0; + + for (int j = 0; j < QK_K/64; ++j) { + const ggml_uint8x16x2_t q4bits = ggml_vld1q_u8_x2(q4); q4 += 32; + + q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; + q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); + q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); + + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); + sumi1 += vaddvq_s32(p1) * scales[2*j+0]; + + q8bytes = ggml_vld1q_s8_x2(q8); q8 += 32; + q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); + q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); + + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); + + sumi2 += vaddvq_s32(p2) * scales[2*j+1]; + } + + sumf += d * (sumi1 + sumi2); + + } + + *s = sumf; + +#elif defined __AVX2__ + + const __m256i m4 = _mm256_set1_epi8(0xF); + + __m256 acc = _mm256_setzero_ps(); + __m128 acc_m = _mm_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); + + const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); + const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); + const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); + acc_m = _mm_fmadd_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod), acc_m); + + const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); + const __m256i scales = MM256_SET_M128I(sc128, sc128); + + __m256i sumi = _mm256_setzero_si256(); + + for (int j = 0; j < QK_K/64; ++j) { + + const __m256i scale_l = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); + const __m256i scale_h = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); + + const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; + const __m256i q4l = _mm256_and_si256(q4bits, m4); + const __m256i q4h = _mm256_and_si256(_mm256_srli_epi16(q4bits, 4), m4); + + const __m256i q8l = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + __m256i p16l = _mm256_maddubs_epi16(q4l, q8l); + p16l = _mm256_madd_epi16(scale_l, p16l); + + const __m256i q8h = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + __m256i p16h = _mm256_maddubs_epi16(q4h, q8h); + p16h = _mm256_madd_epi16(scale_h, p16h); + const __m256i sumj = _mm256_add_epi32(p16l, p16h); + + sumi = _mm256_add_epi32(sumi, sumj); + } + + __m256 vd = _mm256_set1_ps(d); + acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); + + } + + acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); + acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); + + *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); + +#elif defined __AVX__ + + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i m2 = _mm_set1_epi8(0x2); + + __m256 acc = _mm256_setzero_ps(); + __m128 acc_m = _mm_setzero_ps(); + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); + const __m128i scales = _mm_cvtepu8_epi16(utmps); + const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); + + const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); + const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); + const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); + const __m128i prod = _mm_madd_epi16(mins, q8s); + acc_m = _mm_add_ps(_mm_mul_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod)), acc_m); + + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + __m128i shuffle = _mm_set1_epi16(0x0100); + for (int j = 0; j < QK_K/64; ++j) { + + const __m128i scale_l = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + const __m128i scale_h = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + + __m128i q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4l_0 = _mm_and_si128(q4bits, m4); + const __m128i q4h_0 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); + q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4l_1 = _mm_and_si128(q4bits, m4); + const __m128i q4h_1 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); + + const __m128i q8l_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16l = _mm_maddubs_epi16(q4l_0, q8l_0); + p16l = _mm_madd_epi16(scale_l, p16l); + sumi_0 = _mm_add_epi32(sumi_0, p16l); + const __m128i q8l_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + p16l = _mm_maddubs_epi16(q4l_1, q8l_1); + p16l = _mm_madd_epi16(scale_l, p16l); + sumi_1 = _mm_add_epi32(sumi_1, p16l); + + const __m128i q8h_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16h = _mm_maddubs_epi16(q4h_0, q8h_0); + p16h = _mm_madd_epi16(scale_h, p16h); + sumi_0 = _mm_add_epi32(sumi_0, p16h); + const __m128i q8h_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + p16h = _mm_maddubs_epi16(q4h_1, q8h_1); + p16h = _mm_madd_epi16(scale_h, p16h); + sumi_1 = _mm_add_epi32(sumi_1, p16h); + + } + + __m256 vd = _mm256_set1_ps(d); + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); + + } + + acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); + acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); + + *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); + +#elif defined __riscv_v_intrinsic + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + size_t vl = 8; + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); + vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); + vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); + vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); + vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); + + vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); + sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + vl = 32; + + int32_t sum_1 = 0; + int32_t sum_2 = 0; + + vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); + + for (int j = 0; j < QK_K/64; ++j) { + // load Q4 + vuint8m1_t q4_x = __riscv_vle8_v_u8m1(q4, vl); + + // load Q8 and multiply it with lower Q4 nibble + vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); + vint8m1_t q4_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q4_x, 0x0F, vl)); + vint16m2_t qv_0 = __riscv_vwmul_vv_i16m2(q4_0, q8_0, vl); + vint16m1_t vs_0 = __riscv_vredsum_vs_i16m2_i16m1(qv_0, vzero, vl); + + sum_1 += __riscv_vmv_x_s_i16m1_i16(vs_0) * scales[2*j+0]; + + // load Q8 and multiply it with upper Q4 nibble + vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); + vint8m1_t q4_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q4_x, 0x04, vl)); + vint16m2_t qv_1 = __riscv_vwmul_vv_i16m2(q4_1, q8_1, vl); + vint16m1_t vs_1 = __riscv_vredsum_vs_i16m2_i16m1(qv_1, vzero, vl); + + sum_2 += __riscv_vmv_x_s_i16m1_i16(vs_1) * scales[2*j+1]; + + q4 += 32; q8 += 64; + + } + + sumf += d*(sum_1 + sum_2); + + } + + *s = sumf; + +#else + + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + for (int j = 0; j < QK_K/64; ++j) { + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); + a += 32; + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); + a += 32; q4 += 32; + } + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + int sumi = 0; + for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; + a = aux8; + int is = 0; + for (int j = 0; j < QK_K/32; ++j) { + int32_t scale = scales[is++]; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + } + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + sumf -= dmin * sumi; + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; +#endif +} +#else +void ggml_vec_dot_q4_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q4_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#ifdef __ARM_NEON + const uint8x16_t m4b = vdupq_n_u8(0xf); + + const int32x4_t mzero = vdupq_n_s32(0); + + float sumf = 0; + + ggml_int8x16x2_t q4bytes; + ggml_int8x16x4_t q8bytes; + + float sum_mins = 0.f; + + uint16_t aux16[2]; + const uint8_t * restrict scales = (const uint8_t *)aux16; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const uint16_t * restrict a = (const uint16_t *)x[i].scales; + aux16[0] = a[0] & 0x0f0f; + aux16[1] = (a[0] >> 4) & 0x0f0f; + + const int32_t summi = scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3]); + sum_mins += y[i].d * GGML_FP16_TO_FP32(x[i].d[1]) * summi; + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d[0]); + + const ggml_uint8x16x2_t q4bits = ggml_vld1q_u8_x2(q4); + + q8bytes = ggml_vld1q_s8_x4(q8); + q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); + q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); + + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); + const int32_t sumi1 = vaddvq_s32(p1) * scales[0]; + + q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); + q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); + + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[2]), q4bytes.val[1], q8bytes.val[3]); + const int32_t sumi2 = vaddvq_s32(p2) * scales[1]; + + sumf += d * (sumi1 + sumi2); + } + + *s = sumf - sum_mins; + +#elif defined __AVX2__ + + const __m256i m4 = _mm256_set1_epi8(0xF); + + __m256 acc = _mm256_setzero_ps(); + + float summs = 0; + + uint16_t aux16[2]; + const uint8_t * scales = (const uint8_t *)aux16; + + for (int i = 0; i < nb; ++i) { + + const float d = GGML_FP16_TO_FP32(x[i].d[0]) * y[i].d; + const float m = GGML_FP16_TO_FP32(x[i].d[1]) * y[i].d; + const __m256 vd = _mm256_set1_ps(d); + + const uint16_t * a = (const uint16_t *)x[i].scales; + aux16[0] = a[0] & 0x0f0f; + aux16[1] = (a[0] >> 4) & 0x0f0f; + + summs += m * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); + const __m256i q4l = _mm256_and_si256(q4bits, m4); + const __m256i q4h = _mm256_and_si256(_mm256_srli_epi16(q4bits, 4), m4); + + const __m256i q8l = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8h = _mm256_loadu_si256((const __m256i*)(q8+32)); + + const __m256i p16l = _mm256_maddubs_epi16(q4l, q8l); + const __m256i p16h = _mm256_maddubs_epi16(q4h, q8h); + + const __m256i p32l = _mm256_madd_epi16(_mm256_set1_epi16(scales[0]), p16l); + acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(p32l), acc); + + const __m256i p32h = _mm256_madd_epi16(_mm256_set1_epi16(scales[1]), p16h); + acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(p32h), acc); + + } + + *s = hsum_float_8(acc) - summs; + +#elif defined __AVX__ + + const __m128i m4 = _mm_set1_epi8(0xF); + + __m256 acc = _mm256_setzero_ps(); + + float summs = 0; + + uint16_t aux16[2]; + const uint8_t * scales = (const uint8_t *)aux16; + + for (int i = 0; i < nb; ++i) { + + const float d = GGML_FP16_TO_FP32(x[i].d[0]) * y[i].d; + const float m = GGML_FP16_TO_FP32(x[i].d[1]) * y[i].d; + const __m256 vd = _mm256_set1_ps(d); + + const uint16_t * a = (const uint16_t *)x[i].scales; + aux16[0] = a[0] & 0x0f0f; + aux16[1] = (a[0] >> 4) & 0x0f0f; + + summs += m * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); + const __m128i q4bits_0 = _mm256_extractf128_si256(q4bits, 0); + const __m128i q4bits_1 = _mm256_extractf128_si256(q4bits, 1); + const __m128i q4_0 = _mm_and_si128(q4bits_0, m4); + const __m128i q4_1 = _mm_and_si128(q4bits_1, m4); + const __m128i q4_2 = _mm_and_si128(_mm_srli_epi16(q4bits_0, 4), m4); + const __m128i q4_3 = _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4); + + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + + const __m128i p16_0 = _mm_maddubs_epi16(q4_0, _mm256_extractf128_si256(q8_0, 0)); + const __m128i p16_1 = _mm_maddubs_epi16(q4_1, _mm256_extractf128_si256(q8_0, 1)); + const __m128i p16_2 = _mm_maddubs_epi16(q4_2, _mm256_extractf128_si256(q8_1, 0)); + const __m128i p16_3 = _mm_maddubs_epi16(q4_3, _mm256_extractf128_si256(q8_1, 1)); + + const __m128i p32_0 = _mm_madd_epi16(_mm_set1_epi16(scales[0]), p16_0); + const __m128i p32_1 = _mm_madd_epi16(_mm_set1_epi16(scales[0]), p16_1); + acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(MM256_SET_M128I(p32_1, p32_0))), acc); + + const __m128i p32_2 = _mm_madd_epi16(_mm_set1_epi16(scales[1]), p16_2); + const __m128i p32_3 = _mm_madd_epi16(_mm_set1_epi16(scales[1]), p16_3); + acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(MM256_SET_M128I(p32_3, p32_2))), acc); + + } + + *s = hsum_float_8(acc) - summs; + +#elif defined __riscv_v_intrinsic + + uint16_t s16[2]; + const uint8_t * restrict scales = (const uint8_t *)s16; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + const uint16_t * restrict b = (const uint16_t *)x[i].scales; + s16[0] = b[0] & 0x0f0f; + s16[1] = (b[0] >> 4) & 0x0f0f; + + sumf -= y[i].d * GGML_FP16_TO_FP32(x[i].d[1]) * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d[0]); + + size_t vl = 32; + + vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); + + // load Q4 + vuint8m1_t q4_x = __riscv_vle8_v_u8m1(q4, vl); + + // load Q8 and multiply it with lower Q4 nibble + vint8m1_t q4_a = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q4_x, 0x0F, vl)); + vint16m2_t va_0 = __riscv_vwmul_vv_i16m2(q4_a, __riscv_vle8_v_i8m1(q8, vl), vl); + vint16m1_t aux1 = __riscv_vredsum_vs_i16m2_i16m1(va_0, vzero, vl); + + sumf += d*scales[0]*__riscv_vmv_x_s_i16m1_i16(aux1); + + // load Q8 and multiply it with upper Q4 nibble + vint8m1_t q4_s = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q4_x, 0x04, vl)); + vint16m2_t va_1 = __riscv_vwmul_vv_i16m2(q4_s, __riscv_vle8_v_i8m1(q8+32, vl), vl); + vint16m1_t aux2 = __riscv_vredsum_vs_i16m2_i16m1(va_1, vzero, vl); + + sumf += d*scales[1]*__riscv_vmv_x_s_i16m1_i16(aux2); + + } + + *s = sumf; + +#else + + uint8_t aux8[QK_K]; + int16_t aux16[16]; + float sums [8]; + memset(sums, 0, 8*sizeof(float)); + + uint16_t s16[2]; + const uint8_t * restrict scales = (const uint8_t *)s16; + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q4 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + uint8_t * restrict a = aux8; + for (int l = 0; l < 32; ++l) a[l+ 0] = q4[l] & 0xF; + for (int l = 0; l < 32; ++l) a[l+32] = q4[l] >> 4; + + const uint16_t * restrict b = (const uint16_t *)x[i].scales; + s16[0] = b[0] & 0x0f0f; + s16[1] = (b[0] >> 4) & 0x0f0f; + + sumf -= y[i].d * GGML_FP16_TO_FP32(x[i].d[1]) * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d[0]); + + for (int j = 0; j < QK_K/32; ++j) { + for (int l = 0; l < 16; ++l) aux16[l] = q8[l] * a[l]; + q8 += 16; a += 16; + for (int l = 0; l < 16; ++l) aux16[l] += q8[l] * a[l]; + q8 += 16; a += 16; + const float dl = d * scales[j]; + for (int l = 0; l < 8; ++l) sums[l] += dl * (aux16[l] + aux16[l+8]); + } + } + for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; +#endif +} +#endif + +#if QK_K == 256 +void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + + const block_q5_K * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + + static const uint32_t kmask1 = 0x3f3f3f3f; + static const uint32_t kmask2 = 0x0f0f0f0f; + static const uint32_t kmask3 = 0x03030303; + + uint32_t utmp[4]; + +#ifdef __ARM_NEON + const uint8x16_t m4b = vdupq_n_u8(0xf); + const uint8x16_t mone = vdupq_n_u8(1); + const uint8x16_t mtwo = vdupq_n_u8(2); + const int32x4_t mzero = vdupq_n_s32(0); + + ggml_int8x16x4_t q5bytes; + + float sumf = 0; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const uint8x8_t mins8 = vld1_u8((const uint8_t*)utmp + 8); + const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(mins8)); + const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), + vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); + int32_t sumi_mins = vaddvq_s32(prod); + + const uint8_t * scales = (const uint8_t *)utmp; + + const uint8_t * restrict q5 = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); + + ggml_uint8x16x4_t q5h; + + int32_t sumi = 0; + + for (int j = 0; j < QK_K/64; ++j) { + + const ggml_uint8x16x2_t q5bits = ggml_vld1q_u8_x2(q5); q5 += 32; + const ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; + + q5h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); + q5h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); + q5h.val[2] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[0]), 3); + q5h.val[3] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[1]), 3); + qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 2); + qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 2); + + q5bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[0], m4b), q5h.val[0])); + q5bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[1], m4b), q5h.val[1])); + q5bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[0], 4), q5h.val[2])); + q5bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[1], 4), q5h.val[3])); + + sumi += vaddvq_s32(ggml_vdotq_s32(ggml_vdotq_s32(mzero, q5bytes.val[0], q8bytes.val[0]), q5bytes.val[1], q8bytes.val[1])) * *scales++; + sumi += vaddvq_s32(ggml_vdotq_s32(ggml_vdotq_s32(mzero, q5bytes.val[2], q8bytes.val[2]), q5bytes.val[3], q8bytes.val[3])) * *scales++; + } + + sumf += d * sumi - dmin * sumi_mins; + } + + *s = sumf; + +#elif defined __AVX2__ + + const __m256i m4 = _mm256_set1_epi8(0xF); + const __m128i mzero = _mm_setzero_si128(); + const __m256i mone = _mm256_set1_epi8(1); + + __m256 acc = _mm256_setzero_ps(); + + float summs = 0.f; + + for (int i = 0; i < nb; ++i) { + + const uint8_t * restrict q5 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + +#if QK_K == 256 + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; +#else + // TODO + const float d = 0, dmin = 0; +#endif + + const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); + + const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); + const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); + const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); + const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); + summs += dmin * _mm_extract_epi32(hsum, 0); + + const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); + const __m256i scales = MM256_SET_M128I(sc128, sc128); + + const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].qh); + __m256i hmask = mone; + + __m256i sumi = _mm256_setzero_si256(); + + int bit = 0; + + for (int j = 0; j < QK_K/64; ++j) { + + const __m256i scale_0 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); + const __m256i scale_1 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); + + const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); q5 += 32; + + const __m256i q5l_0 = _mm256_and_si256(q5bits, m4); + const __m256i q5h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); + const __m256i q5_0 = _mm256_add_epi8(q5l_0, q5h_0); + hmask = _mm256_slli_epi16(hmask, 1); + + const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4); + const __m256i q5h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); + const __m256i q5_1 = _mm256_add_epi8(q5l_1, q5h_1); + hmask = _mm256_slli_epi16(hmask, 1); + + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + __m256i p16_0 = _mm256_maddubs_epi16(q5_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q5_1, q8_1); + + p16_0 = _mm256_madd_epi16(scale_0, p16_0); + p16_1 = _mm256_madd_epi16(scale_1, p16_1); + + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); + + } + + __m256 vd = _mm256_set1_ps(d); + acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); + + } + + *s = hsum_float_8(acc) + summs; + +#elif defined __AVX__ + + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i mzero = _mm_setzero_si128(); + const __m128i mone = _mm_set1_epi8(1); + const __m128i m2 = _mm_set1_epi8(2); + + __m256 acc = _mm256_setzero_ps(); + + float summs = 0.f; + + for (int i = 0; i < nb; ++i) { + + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + + const uint8_t * restrict q5 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); + const __m128i scales = _mm_cvtepu8_epi16(utmps); + const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); + + const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); + const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); + const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); + const __m128i prod = _mm_madd_epi16(mins, q8s); + const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); + summs += dmin * _mm_extract_epi32(hsum, 0); + + const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].qh[0]); + const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].qh[16]); + __m128i hmask = mone; + + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); + + int bit = 0; + + __m128i shuffle = _mm_set1_epi16(0x0100); + for (int j = 0; j < QK_K/64; ++j) { + + const __m128i scale_0 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + const __m128i scale_1 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi16(shuffle, m2); + + const __m128i q5bits_0 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; + const __m128i q5bits_1 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; + + __m128i q5l_0 = _mm_and_si128(q5bits_0, m4); + __m128i q5l_1 = _mm_and_si128(q5bits_1, m4); + __m128i q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); + __m128i q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); + __m128i q5_0 = _mm_add_epi8(q5l_0, q5h_0); + __m128i q5_1 = _mm_add_epi8(q5l_1, q5h_1); + hmask = _mm_slli_epi16(hmask, 1); + + __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16_0 = _mm_maddubs_epi16(q5_0, q8_0); + __m128i p16_1 = _mm_maddubs_epi16(q5_1, q8_1); + p16_0 = _mm_madd_epi16(scale_0, p16_0); + p16_1 = _mm_madd_epi16(scale_0, p16_1); + + q5l_0 = _mm_and_si128(_mm_srli_epi16(q5bits_0, 4), m4); + q5l_1 = _mm_and_si128(_mm_srli_epi16(q5bits_1, 4), m4); + q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); + q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); + q5_0 = _mm_add_epi8(q5l_0, q5h_0); + q5_1 = _mm_add_epi8(q5l_1, q5h_1); + hmask = _mm_slli_epi16(hmask, 1); + + q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + __m128i p16_2 = _mm_maddubs_epi16(q5_0, q8_0); + __m128i p16_3 = _mm_maddubs_epi16(q5_1, q8_1); + p16_2 = _mm_madd_epi16(scale_1, p16_2); + p16_3 = _mm_madd_epi16(scale_1, p16_3); + + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); + + } + + __m256 vd = _mm256_set1_ps(d); + __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); + acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); + + } + + *s = hsum_float_8(acc) + summs; + +#elif defined __riscv_v_intrinsic + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + float sumf = 0; + float sums = 0.0; + + size_t vl; + + for (int i = 0; i < nb; ++i) { + + vl = 8; + + const uint8_t * restrict q5 = x[i].qs; + const uint8_t * restrict hm = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + + vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); + vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); + vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); + + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; + + vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); + vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); + vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); + + vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); + sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); + + vl = 32; + int32_t aux32 = 0; + int is = 0; + + uint8_t m = 1; + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + vuint8m1_t vqh = __riscv_vle8_v_u8m1(hm, vl); + + for (int j = 0; j < QK_K/64; ++j) { + // load Q5 and Q8 + vuint8m1_t q5_x = __riscv_vle8_v_u8m1(q5, vl); + vint8m1_t q8_y1 = __riscv_vle8_v_i8m1(q8, vl); + vint8m1_t q8_y2 = __riscv_vle8_v_i8m1(q8+32, vl); + + // compute mask for addition + vint8m1_t q5_a = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q5_x, 0x0F, vl)); + vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_1 = __riscv_vmsne_vx_u8m1_b8(qh_m1, 0, vl); + vint8m1_t q5_m1 = __riscv_vadd_vx_i8m1_m(vmask_1, q5_a, 16, vl); + m <<= 1; + + vint8m1_t q5_l = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q5_x, 0x04, vl)); + vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); + vbool8_t vmask_2 = __riscv_vmsne_vx_u8m1_b8(qh_m2, 0, vl); + vint8m1_t q5_m2 = __riscv_vadd_vx_i8m1_m(vmask_2, q5_l, 16, vl); + m <<= 1; + + vint16m2_t v0 = __riscv_vwmul_vv_i16m2(q5_m1, q8_y1, vl); + vint16m2_t v1 = __riscv_vwmul_vv_i16m2(q5_m2, q8_y2, vl); + + vint32m4_t vs1 = __riscv_vwmul_vx_i32m4(v0, scales[is++], vl); + vint32m4_t vs2 = __riscv_vwmul_vx_i32m4(v1, scales[is++], vl); + + vint32m1_t vacc1 = __riscv_vredsum_vs_i32m4_i32m1(vs1, vzero, vl); + vint32m1_t vacc2 = __riscv_vredsum_vs_i32m4_i32m1(vs2, vzero, vl); + + aux32 += __riscv_vmv_x_s_i32m1_i32(vacc1) + __riscv_vmv_x_s_i32m1_i32(vacc2); + q5 += 32; q8 += 64; + + } + + vfloat32m1_t vaux = __riscv_vfmul_vf_f32m1(__riscv_vfmv_v_f_f32m1(aux32, 1), d, 1); + sums += __riscv_vfmv_f_s_f32m1_f32(vaux); + + } + + *s = sumf+sums; + +#else + + const uint8_t * scales = (const uint8_t*)&utmp[0]; + const uint8_t * mins = (const uint8_t*)&utmp[2]; + + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + float sumf = 0; for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict hm = x[i].hmask; + const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict hm = x[i].qh; const int8_t * restrict q8 = y[i].qs; memset(aux32, 0, 8*sizeof(int32_t)); int8_t * restrict a = aux8; uint8_t m = 1; - for (int j = 0; j < QK_K; j += 128) { - for (int l = 0; l < 32; ++l) a[l] = q3[l] & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 2) & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 4) & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + for (int j = 0; j < QK_K/64; ++j) { + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); + for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (q3[l] >> 6) & 3; - for (int l = 0; l < 32; ++l) a[l] -= (hm[l] & m ? 0 : 4); + for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); + for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); a += 32; m <<= 1; - q3 += 32; + q4 += 32; } - a = aux8; + memcpy(utmp, x[i].scales, 12); + utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); + const uint32_t uaux = utmp[1] & kmask1; + utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); + utmp[2] = uaux; + utmp[0] &= kmask1; - memcpy(auxs, x[i].scales, 12); - uint32_t tmp = auxs[2]; - auxs[2] = ((auxs[0] >> 4) & kmask2) | (((tmp >> 4) & kmask1) << 4); - auxs[3] = ((auxs[1] >> 4) & kmask2) | (((tmp >> 6) & kmask1) << 4); - auxs[0] = (auxs[0] & kmask2) | (((tmp >> 0) & kmask1) << 4); - auxs[1] = (auxs[1] & kmask2) | (((tmp >> 2) & kmask1) << 4); - for (int j = 0; j < QK_K/16; ++j) { + int sumi = 0; + for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; + a = aux8; + int is = 0; + for (int j = 0; j < QK_K/32; ++j) { + int32_t scale = scales[is++]; for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; q8 += 8; a += 8; for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += (scales[j] - 32) * aux16[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; q8 += 8; a += 8; } const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + sumf -= dmin * sumi; } for (int l = 0; l < 8; ++l) sumf += sums[l]; *s = sumf; - #endif - } #else -void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { +void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q3_K * restrict x = vx; + const block_q5_K * restrict x = vx; const block_q8_K * restrict y = vy; const int nb = n / QK_K; #ifdef __ARM_NEON + const uint8x16_t m4b = vdupq_n_u8(0xf); + const uint8x16_t mh = vdupq_n_u8(16); + const int32x4_t mzero = vdupq_n_s32(0); -#ifdef __ARM_FEATURE_DOTPROD - const int32x4_t vzero = vdupq_n_s32(0); -#endif - - const uint8x16_t m3b = vdupq_n_u8(0x3); - const uint8x16_t mh = vdupq_n_u8(4); - - int8x16x4_t q3bytes; - - uint16_t aux16[2]; - int8_t * scales = (int8_t *)aux16; + ggml_int8x16x4_t q5bytes; + ggml_uint8x16x4_t q5h; - float sum = 0; + float sumf = 0; for (int i = 0; i < nb; ++i) { - uint8x16x4_t q3h; - - const uint8x8_t hbits = vld1_u8(x[i].hmask); - const uint8x16_t q3bits = vld1q_u8(x[i].qs); - const int8x16x4_t q8bytes = vld1q_s8_x4(y[i].qs); - - const uint16_t a = *(const uint16_t *)x[i].scales; - aux16[0] = a & 0x0f0f; - aux16[1] = (a >> 4) & 0x0f0f; - - for (int j = 0; j < 4; ++j) scales[j] -= 8; + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const int8_t * sc = x[i].scales; - int32_t isum = -4*(scales[0] * y[i].bsums[0] + scales[2] * y[i].bsums[1] + scales[1] * y[i].bsums[2] + scales[3] * y[i].bsums[3]); + const uint8_t * restrict q5 = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; - const float d = y[i].d * (float)x[i].d; + const uint8x8_t qhbits = vld1_u8(qh); - const uint8x16_t htmp = vcombine_u8(hbits, vshr_n_u8(hbits, 1)); - q3h.val[0] = vandq_u8(mh, vshlq_n_u8(htmp, 2)); - q3h.val[1] = vandq_u8(mh, htmp); - q3h.val[2] = vandq_u8(mh, vshrq_n_u8(htmp, 2)); - q3h.val[3] = vandq_u8(mh, vshrq_n_u8(htmp, 4)); + const ggml_uint8x16x2_t q5bits = ggml_vld1q_u8_x2(q5); + const ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); - q3bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q3bits, m3b), q3h.val[0])); - q3bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(vshrq_n_u8(q3bits, 2), m3b), q3h.val[1])); - q3bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(vshrq_n_u8(q3bits, 4), m3b), q3h.val[2])); - q3bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q3bits, 6), q3h.val[3])); + const uint8x16_t htmp = vcombine_u8(qhbits, vshr_n_u8(qhbits, 1)); + q5h.val[0] = vbicq_u8(mh, vshlq_n_u8(htmp, 4)); + q5h.val[1] = vbicq_u8(mh, vshlq_n_u8(htmp, 2)); + q5h.val[2] = vbicq_u8(mh, htmp); + q5h.val[3] = vbicq_u8(mh, vshrq_n_u8(htmp, 2)); -#if defined(__ARM_FEATURE_DOTPROD) - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[0], q8bytes.val[0])) * scales[0]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[1], q8bytes.val[1])) * scales[2]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[2], q8bytes.val[2])) * scales[1]; - isum += vaddvq_s32(vdotq_s32(vzero, q3bytes.val[3], q8bytes.val[3])) * scales[3]; -#else - const int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q3bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q3bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q3bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - const int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q3bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q3bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - isum += vaddvq_s16(p0) * scales[0] + vaddvq_s16(p1) * scales[2] + vaddvq_s16(p2) * scales[1] + vaddvq_s16(p3) * scales[3]; -#endif + q5bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q5bits.val[0], m4b)), vreinterpretq_s8_u8(q5h.val[0])); + q5bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q5bits.val[1], m4b)), vreinterpretq_s8_u8(q5h.val[1])); + q5bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vshrq_n_u8(q5bits.val[0], 4)), vreinterpretq_s8_u8(q5h.val[2])); + q5bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vshrq_n_u8(q5bits.val[1], 4)), vreinterpretq_s8_u8(q5h.val[3])); - sum += d * isum; + int32_t sumi1 = sc[0] * vaddvq_s32(ggml_vdotq_s32(mzero, q5bytes.val[0], q8bytes.val[0])); + int32_t sumi2 = sc[1] * vaddvq_s32(ggml_vdotq_s32(mzero, q5bytes.val[1], q8bytes.val[1])); + int32_t sumi3 = sc[2] * vaddvq_s32(ggml_vdotq_s32(mzero, q5bytes.val[2], q8bytes.val[2])); + int32_t sumi4 = sc[3] * vaddvq_s32(ggml_vdotq_s32(mzero, q5bytes.val[3], q8bytes.val[3])); + sumf += d * (sumi1 + sumi2 + sumi3 + sumi4); } - *s = sum; + *s = sumf; #elif defined __AVX2__ - const __m256i m3 = _mm256_set1_epi8(3); - const __m256i m1 = _mm256_set1_epi8(1); + const __m256i m4 = _mm256_set1_epi8(0xF); + const __m256i mone = _mm256_set1_epi8(1); __m256 acc = _mm256_setzero_ps(); - uint64_t aux64; - - uint16_t aux16[2]; - const int8_t * aux8 = (const int8_t *)aux16; - for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict q5 = x[i].qs; const int8_t * restrict q8 = y[i].qs; - const uint16_t a = *(const uint16_t *)x[i].scales; - aux16[0] = a & 0x0f0f; - aux16[1] = (a >> 4) & 0x0f0f; + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const __m256i scale_0 = MM256_SET_M128I(_mm_set1_epi16(aux8[2] - 8), _mm_set1_epi16(aux8[0] - 8)); - const __m256i scale_1 = MM256_SET_M128I(_mm_set1_epi16(aux8[3] - 8), _mm_set1_epi16(aux8[1] - 8)); + const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); - memcpy(&aux64, x[i].hmask, 8); + const __m256i scale_l = MM256_SET_M128I(_mm_set1_epi16(x[i].scales[1]), _mm_set1_epi16(x[i].scales[0])); + const __m256i scale_h = MM256_SET_M128I(_mm_set1_epi16(x[i].scales[3]), _mm_set1_epi16(x[i].scales[2])); - const __m128i haux = _mm_set_epi64x(aux64 >> 1, aux64 >> 0); - __m256i q3h_0 = MM256_SET_M128I(_mm_srli_epi16(haux, 2), haux); - __m256i q3h_1 = _mm256_srli_epi16(q3h_0, 4); - q3h_0 = _mm256_slli_epi16(_mm256_andnot_si256(q3h_0, m1), 2); - q3h_1 = _mm256_slli_epi16(_mm256_andnot_si256(q3h_1, m1), 2); + int64_t aux64; + memcpy(&aux64, x[i].qh, 8); + const __m128i haux128 = _mm_set_epi64x(aux64 >> 1, aux64); + const __m256i haux256 = MM256_SET_M128I(_mm_srli_epi16(haux128, 2), haux128); - // load low 2 bits - const __m128i q3bits = _mm_loadu_si128((const __m128i*)q3); + const __m256i q5h_0 = _mm256_slli_epi16(_mm256_andnot_si256(haux256, mone), 4); + const __m256i q5h_1 = _mm256_slli_epi16(_mm256_andnot_si256(_mm256_srli_epi16(haux256, 4), mone), 4); - // prepare low and high bits - const __m256i q3aux = MM256_SET_M128I(_mm_srli_epi16(q3bits, 2), q3bits); - const __m256i q3l_0 = _mm256_and_si256(q3aux, m3); - const __m256i q3l_1 = _mm256_and_si256(_mm256_srli_epi16(q3aux, 4), m3); + const __m256i q5l_0 = _mm256_and_si256(q5bits, m4); + const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4); - // load Q8 quants const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm256_maddubs_epi16, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - const __m256i q8s_0 = _mm256_maddubs_epi16(q3h_0, q8_0); - const __m256i q8s_1 = _mm256_maddubs_epi16(q3h_1, q8_1); - - __m256i p16_0 = _mm256_maddubs_epi16(q3l_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q3l_1, q8_1); - - p16_0 = _mm256_sub_epi16(p16_0, q8s_0); - p16_1 = _mm256_sub_epi16(p16_1, q8s_1); - - // multiply with scales - p16_0 = _mm256_madd_epi16(scale_0, p16_0); - p16_1 = _mm256_madd_epi16(scale_1, p16_1); + const __m256i p16_0 = _mm256_madd_epi16(scale_l, _mm256_maddubs_epi16(q5l_0, q8_0)); + const __m256i p16_1 = _mm256_madd_epi16(scale_h, _mm256_maddubs_epi16(q5l_1, q8_1)); + const __m256i s16_0 = _mm256_madd_epi16(scale_l, _mm256_maddubs_epi16(q5h_0, q8_0)); + const __m256i s16_1 = _mm256_madd_epi16(scale_h, _mm256_maddubs_epi16(q5h_1, q8_1)); - p16_0 = _mm256_add_epi32(p16_0, p16_1); + const __m256i dot = _mm256_sub_epi32(_mm256_add_epi32(p16_0, p16_1), _mm256_add_epi32(s16_0, s16_1)); - // multiply with block scale and accumulate - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(p16_0), acc); + acc = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(dot), acc); } @@ -4904,86 +7685,56 @@ void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restri #elif defined __AVX__ - const __m128i m3 = _mm_set1_epi8(3); - const __m128i m1 = _mm_set1_epi8(1); + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i mone = _mm_set1_epi8(1); __m256 acc = _mm256_setzero_ps(); - uint64_t aux64; - - uint16_t aux16[2]; - const int8_t * aux8 = (const int8_t *)aux16; - for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - - const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict q5 = x[i].qs; const int8_t * restrict q8 = y[i].qs; - const uint16_t a = *(const uint16_t *)x[i].scales; - aux16[0] = a & 0x0f0f; - aux16[1] = (a >> 4) & 0x0f0f; + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const __m128i scale_0 = _mm_set1_epi16(aux8[0] - 8); - const __m128i scale_1 = _mm_set1_epi16(aux8[2] - 8); - const __m128i scale_2 = _mm_set1_epi16(aux8[1] - 8); - const __m128i scale_3 = _mm_set1_epi16(aux8[3] - 8); + const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); - memcpy(&aux64, x[i].hmask, 8); + const __m128i scale_0 = _mm_set1_epi16(x[i].scales[0]); + const __m128i scale_1 = _mm_set1_epi16(x[i].scales[1]); + const __m128i scale_2 = _mm_set1_epi16(x[i].scales[2]); + const __m128i scale_3 = _mm_set1_epi16(x[i].scales[3]); - __m128i q3h_0 = _mm_set_epi64x(aux64 >> 1, aux64 >> 0); - __m128i q3h_1 = _mm_srli_epi16(q3h_0, 2); - __m128i q3h_2 = _mm_srli_epi16(q3h_0, 4); - __m128i q3h_3 = _mm_srli_epi16(q3h_0, 6); - q3h_0 = _mm_slli_epi16(_mm_andnot_si128(q3h_0, m1), 2); - q3h_1 = _mm_slli_epi16(_mm_andnot_si128(q3h_1, m1), 2); - q3h_2 = _mm_slli_epi16(_mm_andnot_si128(q3h_2, m1), 2); - q3h_3 = _mm_slli_epi16(_mm_andnot_si128(q3h_3, m1), 2); + int64_t aux64; + memcpy(&aux64, x[i].qh, 8); + const __m128i haux128_0 = _mm_set_epi64x(aux64 >> 1, aux64); + const __m128i haux128_1 = _mm_srli_epi16(haux128_0, 2); - // load low 2 bits - const __m128i q3bits = _mm_loadu_si128((const __m128i*)q3); + const __m128i q5h_0 = _mm_slli_epi16(_mm_andnot_si128(haux128_0, mone), 4); + const __m128i q5h_1 = _mm_slli_epi16(_mm_andnot_si128(haux128_1, mone), 4); + const __m128i q5h_2 = _mm_slli_epi16(_mm_andnot_si128(_mm_srli_epi16(haux128_0, 4), mone), 4); + const __m128i q5h_3 = _mm_slli_epi16(_mm_andnot_si128(_mm_srli_epi16(haux128_1, 4), mone), 4); - // prepare low and high bits - const __m128i q3l_0 = _mm_and_si128(q3bits, m3); - const __m128i q3l_1 = _mm_and_si128(_mm_srli_epi16(q3bits, 2), m3); - const __m128i q3l_2 = _mm_and_si128(_mm_srli_epi16(q3bits, 4), m3); - const __m128i q3l_3 = _mm_and_si128(_mm_srli_epi16(q3bits, 6), m3); + const __m128i q5l_0 = _mm_and_si128(_mm256_extractf128_si256(q5bits, 0), m4); + const __m128i q5l_1 = _mm_and_si128(_mm256_extractf128_si256(q5bits, 1), m4); + const __m128i q5l_2 = _mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q5bits, 0), 4), m4); + const __m128i q5l_3 = _mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q5bits, 1), 4), m4); - // load Q8 quants const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - // Dot product: we multiply the 2 low bits and 1 high bit part separately, so we can use _mm_maddubs_epi16, - // and then subtract. The high bit part has the 2 already subtracted (and so, it is zero if the high bit was not set, - // and 2 if the high bit was set) - const __m128i q8s_0 = _mm_maddubs_epi16(q3h_0, _mm256_extractf128_si256(q8_0, 0)); - const __m128i q8s_1 = _mm_maddubs_epi16(q3h_1, _mm256_extractf128_si256(q8_0, 1)); - const __m128i q8s_2 = _mm_maddubs_epi16(q3h_2, _mm256_extractf128_si256(q8_1, 0)); - const __m128i q8s_3 = _mm_maddubs_epi16(q3h_3, _mm256_extractf128_si256(q8_1, 1)); - - __m128i p16_0 = _mm_maddubs_epi16(q3l_0, _mm256_extractf128_si256(q8_0, 0)); - __m128i p16_1 = _mm_maddubs_epi16(q3l_1, _mm256_extractf128_si256(q8_0, 1)); - __m128i p16_2 = _mm_maddubs_epi16(q3l_2, _mm256_extractf128_si256(q8_1, 0)); - __m128i p16_3 = _mm_maddubs_epi16(q3l_3, _mm256_extractf128_si256(q8_1, 1)); - - p16_0 = _mm_sub_epi16(p16_0, q8s_0); - p16_1 = _mm_sub_epi16(p16_1, q8s_1); - p16_2 = _mm_sub_epi16(p16_2, q8s_2); - p16_3 = _mm_sub_epi16(p16_3, q8s_3); - - // multiply with scales - p16_0 = _mm_madd_epi16(scale_0, p16_0); - p16_1 = _mm_madd_epi16(scale_1, p16_1); - p16_2 = _mm_madd_epi16(scale_2, p16_2); - p16_3 = _mm_madd_epi16(scale_3, p16_3); + const __m128i p16_0 = _mm_madd_epi16(scale_0, _mm_maddubs_epi16(q5l_0, _mm256_extractf128_si256(q8_0, 0))); + const __m128i p16_1 = _mm_madd_epi16(scale_1, _mm_maddubs_epi16(q5l_1, _mm256_extractf128_si256(q8_0, 1))); + const __m128i p16_2 = _mm_madd_epi16(scale_2, _mm_maddubs_epi16(q5l_2, _mm256_extractf128_si256(q8_1, 0))); + const __m128i p16_3 = _mm_madd_epi16(scale_3, _mm_maddubs_epi16(q5l_3, _mm256_extractf128_si256(q8_1, 1))); + const __m128i s16_0 = _mm_madd_epi16(scale_0, _mm_maddubs_epi16(q5h_0, _mm256_extractf128_si256(q8_0, 0))); + const __m128i s16_1 = _mm_madd_epi16(scale_1, _mm_maddubs_epi16(q5h_1, _mm256_extractf128_si256(q8_0, 1))); + const __m128i s16_2 = _mm_madd_epi16(scale_2, _mm_maddubs_epi16(q5h_2, _mm256_extractf128_si256(q8_1, 0))); + const __m128i s16_3 = _mm_madd_epi16(scale_3, _mm_maddubs_epi16(q5h_3, _mm256_extractf128_si256(q8_1, 1))); - p16_0 = _mm_add_epi32(p16_0, p16_2); - p16_1 = _mm_add_epi32(p16_1, p16_3); - __m256i p16 = MM256_SET_M128I(p16_1, p16_0); + const __m128i dot_0 = _mm_sub_epi32(_mm_add_epi32(p16_0, p16_2), _mm_add_epi32(s16_0, s16_2)); + const __m128i dot_1 = _mm_sub_epi32(_mm_add_epi32(p16_1, p16_3), _mm_add_epi32(s16_1, s16_3)); - // multiply with block scale and accumulate - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(p16)), acc); + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(dot_1, dot_0))), acc); } @@ -4991,72 +7742,69 @@ void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restri #elif defined __riscv_v_intrinsic - uint16_t aux16[2]; - int8_t * scales = (int8_t *)aux16; - float sumf = 0; for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q3 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const uint16_t a = *(const uint16_t *)x[i].scales; - aux16[0] = a & 0x0f0f; - aux16[1] = (a >> 4) & 0x0f0f; - - for (int j = 0; j < 4; ++j) scales[j] -= 8; - - int32_t isum = -4*(scales[0] * y[i].bsums[0] + scales[2] * y[i].bsums[1] + scales[1] * y[i].bsums[2] + scales[3] * y[i].bsums[3]); + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const int8_t * sc = x[i].scales; - const float d = y[i].d * (float)x[i].d; + const uint8_t * restrict q5 = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); // load qh - vuint8mf4_t qh_x1 = __riscv_vle8_v_u8mf4(x[i].hmask, 8); + vuint8mf4_t qh_x1 = __riscv_vle8_v_u8mf4(qh, 8); vuint8mf2_t qh_x2 = __riscv_vlmul_ext_v_u8mf4_u8mf2(__riscv_vsrl_vx_u8mf4(qh_x1, 1, 8)); size_t vl = 16; - // extend and combine both qh_x1 and qh_x2 + // combine both qh_1 and qh_2 vuint8mf2_t qh_x = __riscv_vslideup_vx_u8mf2(__riscv_vlmul_ext_v_u8mf4_u8mf2(qh_x1), qh_x2, vl/2, vl); - vuint8mf2_t qh_0 = __riscv_vand_vx_u8mf2(__riscv_vsll_vx_u8mf2(qh_x, 0x2, vl), 0x4, vl); - vuint8mf2_t qh_1 = __riscv_vand_vx_u8mf2(qh_x, 0x4, vl); - vuint8mf2_t qh_2 = __riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl), 0x4, vl); - vuint8mf2_t qh_3 = __riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(qh_x, 0x4, vl), 0x4, vl); + vuint8mf2_t qh_h0 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(__riscv_vsll_vx_u8mf2(qh_x, 0x4, vl), vl), 16, vl); + vuint8mf2_t qh_h1 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(__riscv_vsll_vx_u8mf2(qh_x, 0x2, vl), vl), 16, vl); + vuint8mf2_t qh_h2 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(qh_x, vl), 16, vl); + vuint8mf2_t qh_h3 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(__riscv_vsrl_vx_u8mf2(qh_x, 0x4, vl), vl), 16, vl); - // load Q3 - vuint8mf2_t q3_x = __riscv_vle8_v_u8mf2(q3, vl); + vint8mf2_t qh_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h0); + vint8mf2_t qh_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h1); + vint8mf2_t qh_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h2); + vint8mf2_t qh_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h3); - vuint8mf2_t q3h_0 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(q3_x, 0x3, vl), qh_0, vl); - vuint8mf2_t q3h_1 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q3_x, 2, vl), 0x3, vl), qh_1, vl); - vuint8mf2_t q3h_2 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(__riscv_vsrl_vx_u8mf2(q3_x, 4, vl), 0x3, vl), qh_2, vl); - vuint8mf2_t q3h_3 = __riscv_vor_vv_u8mf2(__riscv_vsrl_vx_u8mf2(q3_x, 0x6, vl), qh_3, vl); + // load q5 + vuint8mf2_t q5_x1 = __riscv_vle8_v_u8mf2(q5, vl); + vuint8mf2_t q5_x2 = __riscv_vle8_v_u8mf2(q5+16, vl); - vint8mf2_t q3_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_0); - vint8mf2_t q3_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_1); - vint8mf2_t q3_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_2); - vint8mf2_t q3_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(q3h_3); + vint8mf2_t q5s_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(q5_x1, 0xF, vl)); + vint8mf2_t q5s_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(q5_x2, 0xF, vl)); + vint8mf2_t q5s_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vsrl_vx_u8mf2(q5_x1, 0x4, vl)); + vint8mf2_t q5s_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vsrl_vx_u8mf2(q5_x2, 0x4, vl)); - // load Q8 and take product with Q3 - vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q3_0, __riscv_vle8_v_i8mf2(q8, vl), vl); - vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q3_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); - vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q3_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); - vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q3_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); + vint8mf2_t q5_0 = __riscv_vsub_vv_i8mf2(q5s_0, qh_0, vl); + vint8mf2_t q5_1 = __riscv_vsub_vv_i8mf2(q5s_1, qh_1, vl); + vint8mf2_t q5_2 = __riscv_vsub_vv_i8mf2(q5s_2, qh_2, vl); + vint8mf2_t q5_3 = __riscv_vsub_vv_i8mf2(q5s_3, qh_3, vl); + + // load Q8 and multiply it with Q5 + vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q5_0, __riscv_vle8_v_i8mf2(q8, vl), vl); + vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q5_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); + vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q5_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); + vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q5_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); vint32m1_t vs_0 = __riscv_vwredsum_vs_i16m1_i32m1(p0, vzero, vl); vint32m1_t vs_1 = __riscv_vwredsum_vs_i16m1_i32m1(p1, vzero, vl); vint32m1_t vs_2 = __riscv_vwredsum_vs_i16m1_i32m1(p2, vzero, vl); vint32m1_t vs_3 = __riscv_vwredsum_vs_i16m1_i32m1(p3, vzero, vl); - isum += __riscv_vmv_x_s_i32m1_i32(vs_0) * scales[0]; - isum += __riscv_vmv_x_s_i32m1_i32(vs_1) * scales[2]; - isum += __riscv_vmv_x_s_i32m1_i32(vs_2) * scales[1]; - isum += __riscv_vmv_x_s_i32m1_i32(vs_3) * scales[3]; + int32_t sumi1 = sc[0] * __riscv_vmv_x_s_i32m1_i32(vs_0); + int32_t sumi2 = sc[1] * __riscv_vmv_x_s_i32m1_i32(vs_1); + int32_t sumi3 = sc[2] * __riscv_vmv_x_s_i32m1_i32(vs_2); + int32_t sumi4 = sc[3] * __riscv_vmv_x_s_i32m1_i32(vs_3); - sumf += d * isum; + sumf += d * (sumi1 + sumi2 + sumi3 + sumi4); } @@ -5064,371 +7812,429 @@ void ggml_vec_dot_q3_K_q8_K(const int n, float * restrict s, const void * restri #else - int8_t aux8[QK_K]; - int16_t aux16[8]; + int8_t aux8[QK_K]; + int16_t aux16[16]; float sums [8]; - int32_t aux32[8]; - int32_t scales[4]; memset(sums, 0, 8*sizeof(float)); float sumf = 0; for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q3 = x[i].qs; - const uint8_t * restrict hm = x[i].hmask; + const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict hm = x[i].qh; const int8_t * restrict q8 = y[i].qs; int8_t * restrict a = aux8; - for (int l = 0; l < 8; ++l) { - a[l+ 0] = (int8_t)((q3[l+0] >> 0) & 3) - (hm[l] & 0x01 ? 0 : 4); - a[l+ 8] = (int8_t)((q3[l+8] >> 0) & 3) - (hm[l] & 0x02 ? 0 : 4); - a[l+16] = (int8_t)((q3[l+0] >> 2) & 3) - (hm[l] & 0x04 ? 0 : 4); - a[l+24] = (int8_t)((q3[l+8] >> 2) & 3) - (hm[l] & 0x08 ? 0 : 4); - a[l+32] = (int8_t)((q3[l+0] >> 4) & 3) - (hm[l] & 0x10 ? 0 : 4); - a[l+40] = (int8_t)((q3[l+8] >> 4) & 3) - (hm[l] & 0x20 ? 0 : 4); - a[l+48] = (int8_t)((q3[l+0] >> 6) & 3) - (hm[l] & 0x40 ? 0 : 4); - a[l+56] = (int8_t)((q3[l+8] >> 6) & 3) - (hm[l] & 0x80 ? 0 : 4); + for (int l = 0; l < 32; ++l) { + a[l+ 0] = q4[l] & 0xF; + a[l+32] = q4[l] >> 4; + } + for (int is = 0; is < 8; ++is) { + uint8_t m = 1 << is; + for (int l = 0; l < 8; ++l) a[8*is + l] -= (hm[l] & m ? 0 : 16); } - scales[0] = (x[i].scales[0] & 0xF) - 8; - scales[1] = (x[i].scales[0] >> 4) - 8; - scales[2] = (x[i].scales[1] & 0xF) - 8; - scales[3] = (x[i].scales[1] >> 4) - 8; + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const int8_t * restrict sc = x[i].scales; - memset(aux32, 0, 8*sizeof(int32_t)); for (int j = 0; j < QK_K/16; ++j) { - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] += q8[l] * a[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux32[l] += scales[j] * aux16[l]; + const float dl = d * sc[j]; + for (int l = 0; l < 16; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) sums[l] += dl * (aux16[l] + aux16[8+l]); + q8 += 16; a += 16; } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; } for (int l = 0; l < 8; ++l) sumf += sums[l]; *s = sumf; - #endif - } #endif + #if QK_K == 256 -void ggml_vec_dot_q4_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { +void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q4_K * restrict x = vx; + const block_q6_K * restrict x = vx; const block_q8_K * restrict y = vy; const int nb = n / QK_K; - static const uint32_t kmask1 = 0x3f3f3f3f; - static const uint32_t kmask2 = 0x0f0f0f0f; - static const uint32_t kmask3 = 0x03030303; - - uint32_t utmp[4]; - #ifdef __ARM_NEON + float sum = 0; - const uint8x16_t m4b = vdupq_n_u8(0xf); -#ifdef __ARM_FEATURE_DOTPROD - const int32x4_t mzero = vdupq_n_s32(0); -#endif + const uint8x16_t m4b = vdupq_n_u8(0xF); + const int32x4_t vzero = vdupq_n_s32(0); + //const int8x16_t m32s = vdupq_n_s8(32); - int8x16x2_t q4bytes; - int8x16x2_t q8bytes; + const uint8x16_t mone = vdupq_n_u8(3); - float sumf = 0; + ggml_int8x16x4_t q6bytes; + ggml_uint8x16x4_t q6h; for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); - - memcpy(utmp, x[i].scales, 12); + const float d_all = GGML_FP16_TO_FP32(x[i].d); - uint32x2_t mins8 = { 0 }; - mins8 = vset_lane_u32(utmp[1] & kmask1, mins8, 0); - mins8 = vset_lane_u32(((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4), mins8, 1); + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[0] &= kmask1; + const int8_t * restrict scale = x[i].scales; - const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(vreinterpret_u8_u32(mins8))); - const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), - vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); - sumf -= dmin * vaddvq_s32(prod); + const ggml_int16x8x2_t q8sums = ggml_vld1q_s16_x2(y[i].bsums); + const int8x16_t scales = vld1q_s8(scale); + const ggml_int16x8x2_t q6scales = {{vmovl_s8(vget_low_s8(scales)), vmovl_s8(vget_high_s8(scales))}}; - const uint8_t * scales = (const uint8_t *)utmp; + const int32x4_t prod = vaddq_s32(vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[0]), vget_low_s16 (q6scales.val[0])), + vmull_s16(vget_high_s16(q8sums.val[0]), vget_high_s16(q6scales.val[0]))), + vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[1]), vget_low_s16 (q6scales.val[1])), + vmull_s16(vget_high_s16(q8sums.val[1]), vget_high_s16(q6scales.val[1])))); + int32_t isum_mins = vaddvq_s32(prod); - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; + int32_t isum = 0; - int32_t sumi1 = 0; - int32_t sumi2 = 0; + for (int j = 0; j < QK_K/128; ++j) { - for (int j = 0; j < QK_K/64; ++j) { + ggml_uint8x16x2_t qhbits = ggml_vld1q_u8_x2(qh); qh += 32; + ggml_uint8x16x4_t q6bits = ggml_vld1q_u8_x4(q6); q6 += 64; + ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; - const uint8x16x2_t q4bits = vld1q_u8_x2(q4); q4 += 32; + q6h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); + q6h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); + uint8x16_t shifted = vshrq_n_u8(qhbits.val[0], 2); + q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[1], 2); + q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); -#ifdef __ARM_FEATURE_DOTPROD - q8bytes = vld1q_s8_x2(q8); q8 += 32; - q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); - q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); + //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])), m32s); + //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])), m32s); + //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])), m32s); + //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])), m32s); + q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])); + q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])); + q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])); + q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])); - const int32x4_t p1 = vdotq_s32(vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); - sumi1 += vaddvq_s32(p1) * scales[2*j+0]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; - q8bytes = vld1q_s8_x2(q8); q8 += 32; - q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); - q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); + scale += 4; - const int32x4_t p2 = vdotq_s32(vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); + q8bytes = ggml_vld1q_s8_x4(q8); q8 += 64; - sumi2 += vaddvq_s32(p2) * scales[2*j+1]; -#else - q8bytes = vld1q_s8_x2(q8); q8 += 32; - q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); - q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); - const int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q4bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q4bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - sumi1 += vaddvq_s16(vaddq_s16(p0, p1)) * scales[2*j+0]; + shifted = vshrq_n_u8(qhbits.val[0], 4); + q6h.val[0] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[1], 4); + q6h.val[1] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[0], 6); + q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits.val[1], 6); + q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - q8bytes = vld1q_s8_x2(q8); q8 += 32; - q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); - q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q4bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q4bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - sumi2 += vaddvq_s16(vaddq_s16(p2, p3)) * scales[2*j+1]; + //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])), m32s); + //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])), m32s); + //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])), m32s); + //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])), m32s); + q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])); + q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])); + q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])); + q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])); -#endif + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; + scale += 4; } - - sumf += d * (sumi1 + sumi2); + //sum += isum * d_all * y[i].d; + sum += d_all * y[i].d * (isum - 32 * isum_mins); } - - *s = sumf; + *s = sum; #elif defined __AVX2__ const __m256i m4 = _mm256_set1_epi8(0xF); + const __m256i m2 = _mm256_set1_epi8(3); + const __m256i m32s = _mm256_set1_epi8(32); __m256 acc = _mm256_setzero_ps(); - __m128 acc_m = _mm_setzero_ps(); - for (int i = 0; i < nb; ++i) { + for (int i = 0; i < nb; ++i) { const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; const int8_t * restrict q8 = y[i].qs; - const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); + const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); + + __m256i sumi = _mm256_setzero_si256(); + + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + const __m128i scale_0 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 0)); + const __m128i scale_1 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 1)); + const __m128i scale_2 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 2)); + const __m128i scale_3 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 3)); + is += 4; + + const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; + const __m256i q4bits2 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; + const __m256i q4bitsH = _mm256_loadu_si256((const __m256i*)qh); qh += 32; - const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); - const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); - const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); - acc_m = _mm_fmadd_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod), acc_m); + const __m256i q4h_0 = _mm256_slli_epi16(_mm256_and_si256(q4bitsH, m2), 4); + const __m256i q4h_1 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 2), m2), 4); + const __m256i q4h_2 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 4), m2), 4); + const __m256i q4h_3 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 6), m2), 4); - const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); - const __m256i scales = MM256_SET_M128I(sc128, sc128); + const __m256i q4_0 = _mm256_or_si256(_mm256_and_si256(q4bits1, m4), q4h_0); + const __m256i q4_1 = _mm256_or_si256(_mm256_and_si256(q4bits2, m4), q4h_1); + const __m256i q4_2 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits1, 4), m4), q4h_2); + const __m256i q4_3 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits2, 4), m4), q4h_3); - __m256i sumi = _mm256_setzero_si256(); + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - for (int j = 0; j < QK_K/64; ++j) { + __m256i q8s_0 = _mm256_maddubs_epi16(m32s, q8_0); + __m256i q8s_1 = _mm256_maddubs_epi16(m32s, q8_1); + __m256i q8s_2 = _mm256_maddubs_epi16(m32s, q8_2); + __m256i q8s_3 = _mm256_maddubs_epi16(m32s, q8_3); - const __m256i scale_l = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); - const __m256i scale_h = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); + __m256i p16_0 = _mm256_maddubs_epi16(q4_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q4_1, q8_1); + __m256i p16_2 = _mm256_maddubs_epi16(q4_2, q8_2); + __m256i p16_3 = _mm256_maddubs_epi16(q4_3, q8_3); - const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; - const __m256i q4l = _mm256_and_si256(q4bits, m4); - const __m256i q4h = _mm256_and_si256(_mm256_srli_epi16(q4bits, 4), m4); + p16_0 = _mm256_sub_epi16(p16_0, q8s_0); + p16_1 = _mm256_sub_epi16(p16_1, q8s_1); + p16_2 = _mm256_sub_epi16(p16_2, q8s_2); + p16_3 = _mm256_sub_epi16(p16_3, q8s_3); - const __m256i q8l = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - __m256i p16l = _mm256_maddubs_epi16(q4l, q8l); - p16l = _mm256_madd_epi16(scale_l, p16l); + p16_0 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_0), p16_0); + p16_1 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_1), p16_1); + p16_2 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_2), p16_2); + p16_3 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_3), p16_3); - const __m256i q8h = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - __m256i p16h = _mm256_maddubs_epi16(q4h, q8h); - p16h = _mm256_madd_epi16(scale_h, p16h); - const __m256i sumj = _mm256_add_epi32(p16l, p16h); + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_2, p16_3)); - sumi = _mm256_add_epi32(sumi, sumj); } - __m256 vd = _mm256_set1_ps(d); - acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); - + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); } - acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); - acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); - - *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); + *s = hsum_float_8(acc); #elif defined __AVX__ const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i m2 = _mm_set1_epi8(0x2); + const __m128i m3 = _mm_set1_epi8(3); + const __m128i m32s = _mm_set1_epi8(32); + const __m128i m2 = _mm_set1_epi8(2); __m256 acc = _mm256_setzero_ps(); - __m128 acc_m = _mm_setzero_ps(); - for (int i = 0; i < nb; ++i) { + for (int i = 0; i < nb; ++i) { const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; const int8_t * restrict q8 = y[i].qs; - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); - const __m128i scales = _mm_cvtepu8_epi16(utmps); - const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); - - const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); - const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); - const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); - const __m128i prod = _mm_madd_epi16(mins, q8s); - acc_m = _mm_add_ps(_mm_mul_ps(_mm_set1_ps(dmin), _mm_cvtepi32_ps(prod)), acc_m); + const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); __m128i sumi_0 = _mm_setzero_si128(); __m128i sumi_1 = _mm_setzero_si128(); - __m128i shuffle = _mm_set1_epi16(0x0100); - for (int j = 0; j < QK_K/64; ++j) { + __m128i shuffle = _mm_set_epi64x(0x0101010101010101, 0x0000000000000000); + for (int j = 0; j < QK_K/128; ++j) { - const __m128i scale_l = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); - const __m128i scale_h = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); + const __m128i q4bitsH_0 = _mm_loadu_si128((const __m128i*)qh); qh += 16; + const __m128i q4bitsH_1 = _mm_loadu_si128((const __m128i*)qh); qh += 16; - __m128i q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4l_0 = _mm_and_si128(q4bits, m4); - const __m128i q4h_0 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); - q4bits = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4l_1 = _mm_and_si128(q4bits, m4); - const __m128i q4h_1 = _mm_and_si128(_mm_srli_epi16(q4bits, 4), m4); + const __m128i q4h_0 = _mm_slli_epi16(_mm_and_si128(q4bitsH_0, m3), 4); + const __m128i q4h_1 = _mm_slli_epi16(_mm_and_si128(q4bitsH_1, m3), 4); + const __m128i q4h_2 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_0, 2), m3), 4); + const __m128i q4h_3 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_1, 2), m3), 4); + const __m128i q4h_4 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_0, 4), m3), 4); + const __m128i q4h_5 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_1, 4), m3), 4); + const __m128i q4h_6 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_0, 6), m3), 4); + const __m128i q4h_7 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_1, 6), m3), 4); - const __m128i q8l_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16l = _mm_maddubs_epi16(q4l_0, q8l_0); - p16l = _mm_madd_epi16(scale_l, p16l); - sumi_0 = _mm_add_epi32(sumi_0, p16l); - const __m128i q8l_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - p16l = _mm_maddubs_epi16(q4l_1, q8l_1); - p16l = _mm_madd_epi16(scale_l, p16l); - sumi_1 = _mm_add_epi32(sumi_1, p16l); + const __m128i q4bits1_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4bits1_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4bits2_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const __m128i q4bits2_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q8h_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16h = _mm_maddubs_epi16(q4h_0, q8h_0); - p16h = _mm_madd_epi16(scale_h, p16h); - sumi_0 = _mm_add_epi32(sumi_0, p16h); - const __m128i q8h_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - p16h = _mm_maddubs_epi16(q4h_1, q8h_1); - p16h = _mm_madd_epi16(scale_h, p16h); - sumi_1 = _mm_add_epi32(sumi_1, p16h); + const __m128i q4_0 = _mm_or_si128(_mm_and_si128(q4bits1_0, m4), q4h_0); + const __m128i q4_1 = _mm_or_si128(_mm_and_si128(q4bits1_1, m4), q4h_1); + const __m128i q4_2 = _mm_or_si128(_mm_and_si128(q4bits2_0, m4), q4h_2); + const __m128i q4_3 = _mm_or_si128(_mm_and_si128(q4bits2_1, m4), q4h_3); + const __m128i q4_4 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_0, 4), m4), q4h_4); + const __m128i q4_5 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_1, 4), m4), q4h_5); + const __m128i q4_6 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_0, 4), m4), q4h_6); + const __m128i q4_7 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_1, 4), m4), q4h_7); + + const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + + __m128i q8s_0 = _mm_maddubs_epi16(m32s, q8_0); + __m128i q8s_1 = _mm_maddubs_epi16(m32s, q8_1); + __m128i q8s_2 = _mm_maddubs_epi16(m32s, q8_2); + __m128i q8s_3 = _mm_maddubs_epi16(m32s, q8_3); + __m128i q8s_4 = _mm_maddubs_epi16(m32s, q8_4); + __m128i q8s_5 = _mm_maddubs_epi16(m32s, q8_5); + __m128i q8s_6 = _mm_maddubs_epi16(m32s, q8_6); + __m128i q8s_7 = _mm_maddubs_epi16(m32s, q8_7); + + __m128i p16_0 = _mm_maddubs_epi16(q4_0, q8_0); + __m128i p16_1 = _mm_maddubs_epi16(q4_1, q8_1); + __m128i p16_2 = _mm_maddubs_epi16(q4_2, q8_2); + __m128i p16_3 = _mm_maddubs_epi16(q4_3, q8_3); + __m128i p16_4 = _mm_maddubs_epi16(q4_4, q8_4); + __m128i p16_5 = _mm_maddubs_epi16(q4_5, q8_5); + __m128i p16_6 = _mm_maddubs_epi16(q4_6, q8_6); + __m128i p16_7 = _mm_maddubs_epi16(q4_7, q8_7); + + p16_0 = _mm_sub_epi16(p16_0, q8s_0); + p16_1 = _mm_sub_epi16(p16_1, q8s_1); + p16_2 = _mm_sub_epi16(p16_2, q8s_2); + p16_3 = _mm_sub_epi16(p16_3, q8s_3); + p16_4 = _mm_sub_epi16(p16_4, q8s_4); + p16_5 = _mm_sub_epi16(p16_5, q8s_5); + p16_6 = _mm_sub_epi16(p16_6, q8s_6); + p16_7 = _mm_sub_epi16(p16_7, q8s_7); + + const __m128i scale_0 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi8(shuffle, m2); + const __m128i scale_1 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi8(shuffle, m2); + const __m128i scale_2 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi8(shuffle, m2); + const __m128i scale_3 = _mm_shuffle_epi8(scales, shuffle); + shuffle = _mm_add_epi8(shuffle, m2); + + p16_0 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_0), p16_0); + p16_1 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_0, scale_0)), p16_1); + p16_2 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_1), p16_2); + p16_3 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_1, scale_1)), p16_3); + p16_4 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_2), p16_4); + p16_5 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_2, scale_2)), p16_5); + p16_6 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_3), p16_6); + p16_7 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_3, scale_3)), p16_7); + + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_4, p16_6)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_5, p16_7)); } - __m256 vd = _mm256_set1_ps(d); __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); - + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi)), acc); } - acc_m = _mm_add_ps(acc_m, _mm_movehl_ps(acc_m, acc_m)); - acc_m = _mm_add_ss(acc_m, _mm_movehdup_ps(acc_m)); - - *s = hsum_float_8(acc) + _mm_cvtss_f32(acc_m); + *s = hsum_float_8(acc); #elif defined __riscv_v_intrinsic - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; - float sumf = 0; - for (int i = 0; i < nb; ++i) { - size_t vl = 8; + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; - vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); - vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); - vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); + const int8_t * restrict scale = x[i].scales; - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; + size_t vl; - vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); - vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); - vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); - sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); + int sum_t = 0; + int is = 0; + + for (int j = 0; j < QK_K/128; ++j) { + + vl = 32; + + // load qh + vuint8m1_t qh_x = __riscv_vle8_v_u8m1(qh, vl); + + // load Q6 + vuint8m1_t q6_0 = __riscv_vle8_v_u8m1(q6, vl); + vuint8m1_t q6_1 = __riscv_vle8_v_u8m1(q6+32, vl); - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; + vuint8m1_t q6a_0 = __riscv_vand_vx_u8m1(q6_0, 0x0F, vl); + vuint8m1_t q6a_1 = __riscv_vand_vx_u8m1(q6_1, 0x0F, vl); + vuint8m1_t q6s_0 = __riscv_vsrl_vx_u8m1(q6_0, 0x04, vl); + vuint8m1_t q6s_1 = __riscv_vsrl_vx_u8m1(q6_1, 0x04, vl); - vl = 32; + vuint8m1_t qh_0 = __riscv_vand_vx_u8m1(qh_x, 0x03, vl); + vuint8m1_t qh_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x2, vl), 0x03 , vl); + vuint8m1_t qh_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x4, vl), 0x03 , vl); + vuint8m1_t qh_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x6, vl), 0x03 , vl); - int32_t sum_1 = 0; - int32_t sum_2 = 0; + vuint8m1_t qhi_0 = __riscv_vor_vv_u8m1(q6a_0, __riscv_vsll_vx_u8m1(qh_0, 0x04, vl), vl); + vuint8m1_t qhi_1 = __riscv_vor_vv_u8m1(q6a_1, __riscv_vsll_vx_u8m1(qh_1, 0x04, vl), vl); + vuint8m1_t qhi_2 = __riscv_vor_vv_u8m1(q6s_0, __riscv_vsll_vx_u8m1(qh_2, 0x04, vl), vl); + vuint8m1_t qhi_3 = __riscv_vor_vv_u8m1(q6s_1, __riscv_vsll_vx_u8m1(qh_3, 0x04, vl), vl); - vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); + vint8m1_t a_0 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_0), 32, vl); + vint8m1_t a_1 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_1), 32, vl); + vint8m1_t a_2 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_2), 32, vl); + vint8m1_t a_3 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_3), 32, vl); - for (int j = 0; j < QK_K/64; ++j) { - // load Q4 - vuint8m1_t q4_x = __riscv_vle8_v_u8m1(q4, vl); + // load Q8 and take product + vint16m2_t va_q_0 = __riscv_vwmul_vv_i16m2(a_0, __riscv_vle8_v_i8m1(q8, vl), vl); + vint16m2_t va_q_1 = __riscv_vwmul_vv_i16m2(a_1, __riscv_vle8_v_i8m1(q8+32, vl), vl); + vint16m2_t va_q_2 = __riscv_vwmul_vv_i16m2(a_2, __riscv_vle8_v_i8m1(q8+64, vl), vl); + vint16m2_t va_q_3 = __riscv_vwmul_vv_i16m2(a_3, __riscv_vle8_v_i8m1(q8+96, vl), vl); - // load Q8 and multiply it with lower Q4 nibble - vint8m1_t q8_0 = __riscv_vle8_v_i8m1(q8, vl); - vint8m1_t q4_0 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q4_x, 0x0F, vl)); - vint16m2_t qv_0 = __riscv_vwmul_vv_i16m2(q4_0, q8_0, vl); - vint16m1_t vs_0 = __riscv_vredsum_vs_i16m2_i16m1(qv_0, vzero, vl); + vl = 16; - sum_1 += __riscv_vmv_x_s_i16m1_i16(vs_0) * scales[2*j+0]; + vint32m2_t vaux_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 0), scale[is+0], vl); + vint32m2_t vaux_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 1), scale[is+1], vl); + vint32m2_t vaux_2 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 0), scale[is+2], vl); + vint32m2_t vaux_3 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 1), scale[is+3], vl); + vint32m2_t vaux_4 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 0), scale[is+4], vl); + vint32m2_t vaux_5 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 1), scale[is+5], vl); + vint32m2_t vaux_6 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 0), scale[is+6], vl); + vint32m2_t vaux_7 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 1), scale[is+7], vl); - // load Q8 and multiply it with upper Q4 nibble - vint8m1_t q8_1 = __riscv_vle8_v_i8m1(q8+32, vl); - vint8m1_t q4_1 = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q4_x, 0x04, vl)); - vint16m2_t qv_1 = __riscv_vwmul_vv_i16m2(q4_1, q8_1, vl); - vint16m1_t vs_1 = __riscv_vredsum_vs_i16m2_i16m1(qv_1, vzero, vl); + vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_0, vaux_1, vl), vzero, vl); + vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_2, vaux_3, vl), isum0, vl); + vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_4, vaux_5, vl), isum1, vl); + vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_6, vaux_7, vl), isum2, vl); - sum_2 += __riscv_vmv_x_s_i16m1_i16(vs_1) * scales[2*j+1]; + sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); - q4 += 32; q8 += 64; + q6 += 64; qh += 32; q8 += 128; is=8; } - sumf += d*(sum_1 + sum_2); + sumf += d * sum_t; } @@ -5436,10 +8242,6 @@ void ggml_vec_dot_q4_K_q8_K(const int n, float * restrict s, const void * restri #else - - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; - int8_t aux8[QK_K]; int16_t aux16[8]; float sums [8]; @@ -5448,35 +8250,26 @@ void ggml_vec_dot_q4_K_q8_K(const int n, float * restrict s, const void * restri float sumf = 0; for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; const int8_t * restrict q8 = y[i].qs; memset(aux32, 0, 8*sizeof(int32_t)); int8_t * restrict a = aux8; - for (int j = 0; j < QK_K/64; ++j) { - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); - a += 32; - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); - a += 32; q4 += 32; + for (int j = 0; j < QK_K; j += 128) { + for (int l = 0; l < 32; ++l) { + a[l + 0] = (int8_t)((q4[l + 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32; + a[l + 32] = (int8_t)((q4[l + 32] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32; + a[l + 64] = (int8_t)((q4[l + 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32; + a[l + 96] = (int8_t)((q4[l + 32] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32; + } + a += 128; + q4 += 64; + qh += 32; } - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - - int sumi = 0; - for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; a = aux8; int is = 0; - for (int j = 0; j < QK_K/32; ++j) { - int32_t scale = scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; + for (int j = 0; j < QK_K/16; ++j) { + int scale = x[i].scales[is++]; for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; q8 += 8; a += 8; @@ -5486,1797 +8279,4400 @@ void ggml_vec_dot_q4_K_q8_K(const int n, float * restrict s, const void * restri } const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; - const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; - sumf -= dmin * sumi; } for (int l = 0; l < 8; ++l) sumf += sums[l]; *s = sumf; #endif } + #else -void ggml_vec_dot_q4_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { + +void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q4_K * restrict x = vx; + const block_q6_K * restrict x = vx; const block_q8_K * restrict y = vy; const int nb = n / QK_K; #ifdef __ARM_NEON + float sum = 0; - const uint8x16_t m4b = vdupq_n_u8(0xf); - -#ifdef __ARM_FEATURE_DOTPROD - const int32x4_t mzero = vdupq_n_s32(0); -#endif - - float sumf = 0; - - int8x16x2_t q4bytes; - int8x16x4_t q8bytes; + const uint8x16_t m4b = vdupq_n_u8(0xF); + const int8x16_t m32s = vdupq_n_s8(32); + const int32x4_t vzero = vdupq_n_s32(0); - float sum_mins = 0.f; + const uint8x16_t mone = vdupq_n_u8(3); - uint16_t aux16[2]; - const uint8_t * restrict scales = (const uint8_t *)aux16; + ggml_int8x16x4_t q6bytes; + ggml_uint8x16x4_t q6h; for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const uint16_t * restrict a = (const uint16_t *)x[i].scales; - aux16[0] = a[0] & 0x0f0f; - aux16[1] = (a[0] >> 4) & 0x0f0f; - - const int32_t summi = scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3]); - sum_mins += y[i].d * (float)x[i].d[1] * summi; - - const float d = y[i].d * (float)x[i].d[0]; + const float d_all = GGML_FP16_TO_FP32(x[i].d); - const uint8x16x2_t q4bits = vld1q_u8_x2(q4); + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; -#ifdef __ARM_FEATURE_DOTPROD - q8bytes = vld1q_s8_x4(q8); - q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); - q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); + const int8_t * restrict scale = x[i].scales; - const int32x4_t p1 = vdotq_s32(vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[0]), q4bytes.val[1], q8bytes.val[1]); - const int32_t sumi1 = vaddvq_s32(p1) * scales[0]; + int32_t isum = 0; - q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); - q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); + uint8x16_t qhbits = vld1q_u8(qh); + ggml_uint8x16x2_t q6bits = ggml_vld1q_u8_x2(q6); + ggml_int8x16x4_t q8bytes = ggml_vld1q_s8_x4(q8); - const int32x4_t p2 = vdotq_s32(vdotq_s32(mzero, q4bytes.val[0], q8bytes.val[2]), q4bytes.val[1], q8bytes.val[3]); - const int32_t sumi2 = vaddvq_s32(p2) * scales[1]; + q6h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits), 4); + uint8x16_t shifted = vshrq_n_u8(qhbits, 2); + q6h.val[1] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits, 4); + q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + shifted = vshrq_n_u8(qhbits, 6); + q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); -#else - q8bytes = vld1q_s8_x4(q8); - q4bytes.val[0] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[0], m4b)); - q4bytes.val[1] = vreinterpretq_s8_u8(vandq_u8 (q4bits.val[1], m4b)); - const int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q4bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q4bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - int32_t sumi1 = vaddvq_s16(vaddq_s16(p0, p1)) * scales[0]; + q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])), m32s); + q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])), m32s); + q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[2])), m32s); + q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[3])), m32s); - q4bytes.val[0] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[0], 4)); - q4bytes.val[1] = vreinterpretq_s8_u8(vshrq_n_u8(q4bits.val[1], 4)); - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[0]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q4bytes.val[0]), vget_high_s8(q8bytes.val[2]))); - const int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q4bytes.val[1]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q4bytes.val[1]), vget_high_s8(q8bytes.val[3]))); - int32_t sumi2 = vaddvq_s16(vaddq_s16(p2, p3)) * scales[1]; + isum += vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + + vaddvq_s32(ggml_vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; -#endif - sumf += d * (sumi1 + sumi2); + sum += isum * d_all * y[i].d; } - - *s = sumf - sum_mins; + *s = sum; #elif defined __AVX2__ const __m256i m4 = _mm256_set1_epi8(0xF); + const __m256i m2 = _mm256_set1_epi8(3); + const __m256i m32s = _mm256_set1_epi8(32); __m256 acc = _mm256_setzero_ps(); - float summs = 0; - - uint16_t aux16[2]; - const uint8_t * scales = (const uint8_t *)aux16; - - for (int i = 0; i < nb; ++i) { - - const float d = GGML_FP16_TO_FP32(x[i].d[0]) * y[i].d; - const float m = GGML_FP16_TO_FP32(x[i].d[1]) * y[i].d; - const __m256 vd = _mm256_set1_ps(d); - - const uint16_t * a = (const uint16_t *)x[i].scales; - aux16[0] = a[0] & 0x0f0f; - aux16[1] = (a[0] >> 4) & 0x0f0f; - - summs += m * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); - - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - - const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); - const __m256i q4l = _mm256_and_si256(q4bits, m4); - const __m256i q4h = _mm256_and_si256(_mm256_srli_epi16(q4bits, 4), m4); - - const __m256i q8l = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8h = _mm256_loadu_si256((const __m256i*)(q8+32)); - - const __m256i p16l = _mm256_maddubs_epi16(q4l, q8l); - const __m256i p16h = _mm256_maddubs_epi16(q4h, q8h); - - const __m256i p32l = _mm256_madd_epi16(_mm256_set1_epi16(scales[0]), p16l); - acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(p32l), acc); - - const __m256i p32h = _mm256_madd_epi16(_mm256_set1_epi16(scales[1]), p16h); - acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(p32h), acc); - - } - - *s = hsum_float_8(acc) - summs; - -#elif defined __AVX__ - - const __m128i m4 = _mm_set1_epi8(0xF); - - __m256 acc = _mm256_setzero_ps(); - - float summs = 0; - - uint16_t aux16[2]; - const uint8_t * scales = (const uint8_t *)aux16; - for (int i = 0; i < nb; ++i) { - const float d = GGML_FP16_TO_FP32(x[i].d[0]) * y[i].d; - const float m = GGML_FP16_TO_FP32(x[i].d[1]) * y[i].d; - const __m256 vd = _mm256_set1_ps(d); - - const uint16_t * a = (const uint16_t *)x[i].scales; - aux16[0] = a[0] & 0x0f0f; - aux16[1] = (a[0] >> 4) & 0x0f0f; - - summs += m * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const uint8_t * restrict q4 = x[i].qs; + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; const int8_t * restrict q8 = y[i].qs; - const __m256i q4bits = _mm256_loadu_si256((const __m256i*)q4); - const __m128i q4bits_0 = _mm256_extractf128_si256(q4bits, 0); - const __m128i q4bits_1 = _mm256_extractf128_si256(q4bits, 1); - const __m128i q4_0 = _mm_and_si128(q4bits_0, m4); - const __m128i q4_1 = _mm_and_si128(q4bits_1, m4); - const __m128i q4_2 = _mm_and_si128(_mm_srli_epi16(q4bits_0, 4), m4); - const __m128i q4_3 = _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4); - - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - - const __m128i p16_0 = _mm_maddubs_epi16(q4_0, _mm256_extractf128_si256(q8_0, 0)); - const __m128i p16_1 = _mm_maddubs_epi16(q4_1, _mm256_extractf128_si256(q8_0, 1)); - const __m128i p16_2 = _mm_maddubs_epi16(q4_2, _mm256_extractf128_si256(q8_1, 0)); - const __m128i p16_3 = _mm_maddubs_epi16(q4_3, _mm256_extractf128_si256(q8_1, 1)); - - const __m128i p32_0 = _mm_madd_epi16(_mm_set1_epi16(scales[0]), p16_0); - const __m128i p32_1 = _mm_madd_epi16(_mm_set1_epi16(scales[0]), p16_1); - acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(MM256_SET_M128I(p32_1, p32_0))), acc); - - const __m128i p32_2 = _mm_madd_epi16(_mm_set1_epi16(scales[1]), p16_2); - const __m128i p32_3 = _mm_madd_epi16(_mm_set1_epi16(scales[1]), p16_3); - acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(MM256_SET_M128I(p32_3, p32_2))), acc); - - } + const __m64 scales_1 = _mm_set1_pi8(x[i].scales[0]); + const __m64 scales_2 = _mm_set1_pi8(x[i].scales[1]); + const __m64 scales_3 = _mm_set1_pi8(x[i].scales[2]); + const __m64 scales_4 = _mm_set1_pi8(x[i].scales[3]); - *s = hsum_float_8(acc) - summs; + __m256i sumi = _mm256_setzero_si256(); -#elif defined __riscv_v_intrinsic + const __m128i scale_0 = _mm_set_epi64(scales_2, scales_1); + const __m128i scale_1 = _mm_set_epi64(scales_4, scales_3); - uint16_t s16[2]; - const uint8_t * restrict scales = (const uint8_t *)s16; + const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); + const __m128i q4bitsH = _mm_loadu_si128((const __m128i*)qh); - float sumf = 0; + const __m256i q4h_0 = _mm256_slli_epi16(_mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q4bitsH, 2), q4bitsH), m2), 4); + const __m256i q4h_1 = _mm256_slli_epi16(_mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q4bitsH, 6), _mm_srli_epi16(q4bitsH, 4)), m2), 4); - for (int i = 0; i < nb; ++i) { + const __m256i q4_0 = _mm256_or_si256(_mm256_and_si256(q4bits1, m4), q4h_0); + const __m256i q4_1 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits1, 4), m4), q4h_1); - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - const uint16_t * restrict b = (const uint16_t *)x[i].scales; - s16[0] = b[0] & 0x0f0f; - s16[1] = (b[0] >> 4) & 0x0f0f; + __m256i q8s_0 = _mm256_maddubs_epi16(m32s, q8_0); + __m256i q8s_1 = _mm256_maddubs_epi16(m32s, q8_1); - sumf -= y[i].d * GGML_FP16_TO_FP32(x[i].d[1]) * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d[0]); + __m256i p16_0 = _mm256_maddubs_epi16(q4_0, q8_0); + __m256i p16_1 = _mm256_maddubs_epi16(q4_1, q8_1); - size_t vl = 32; + p16_0 = _mm256_sub_epi16(p16_0, q8s_0); + p16_1 = _mm256_sub_epi16(p16_1, q8s_1); - vint16m1_t vzero = __riscv_vmv_v_x_i16m1(0, 1); + p16_0 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_0), p16_0); + p16_1 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_1), p16_1); - // load Q4 - vuint8m1_t q4_x = __riscv_vle8_v_u8m1(q4, vl); + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); - // load Q8 and multiply it with lower Q4 nibble - vint8m1_t q4_a = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q4_x, 0x0F, vl)); - vint16m2_t va_0 = __riscv_vwmul_vv_i16m2(q4_a, __riscv_vle8_v_i8m1(q8, vl), vl); - vint16m1_t aux1 = __riscv_vredsum_vs_i16m2_i16m1(va_0, vzero, vl); + acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + } - sumf += d*scales[0]*__riscv_vmv_x_s_i16m1_i16(aux1); + *s = hsum_float_8(acc); - // load Q8 and multiply it with upper Q4 nibble - vint8m1_t q4_s = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q4_x, 0x04, vl)); - vint16m2_t va_1 = __riscv_vwmul_vv_i16m2(q4_s, __riscv_vle8_v_i8m1(q8+32, vl), vl); - vint16m1_t aux2 = __riscv_vredsum_vs_i16m2_i16m1(va_1, vzero, vl); +#elif defined __AVX__ - sumf += d*scales[1]*__riscv_vmv_x_s_i16m1_i16(aux2); + const __m128i m4 = _mm_set1_epi8(0xF); + const __m128i m2 = _mm_set1_epi8(3); + const __m128i m32s = _mm_set1_epi8(32); - } + __m256 acc = _mm256_setzero_ps(); - *s = sumf; + for (int i = 0; i < nb; ++i) { -#else + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - uint8_t aux8[QK_K]; - int16_t aux16[16]; - float sums [8]; - memset(sums, 0, 8*sizeof(float)); + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; - uint16_t s16[2]; - const uint8_t * restrict scales = (const uint8_t *)s16; + const __m64 scales_1 = _mm_set1_pi8(x[i].scales[0]); + const __m64 scales_2 = _mm_set1_pi8(x[i].scales[1]); + const __m64 scales_3 = _mm_set1_pi8(x[i].scales[2]); + const __m64 scales_4 = _mm_set1_pi8(x[i].scales[3]); - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - uint8_t * restrict a = aux8; - for (int l = 0; l < 32; ++l) a[l+ 0] = q4[l] & 0xF; - for (int l = 0; l < 32; ++l) a[l+32] = q4[l] >> 4; + __m128i sumi_0 = _mm_setzero_si128(); + __m128i sumi_1 = _mm_setzero_si128(); - const uint16_t * restrict b = (const uint16_t *)x[i].scales; - s16[0] = b[0] & 0x0f0f; - s16[1] = (b[0] >> 4) & 0x0f0f; + const __m128i scale_0 = _mm_set_epi64(scales_2, scales_1); + const __m128i scale_1 = _mm_set_epi64(scales_4, scales_3); - sumf -= y[i].d * GGML_FP16_TO_FP32(x[i].d[1]) * (scales[2] * (y[i].bsums[0] + y[i].bsums[1]) + scales[3] * (y[i].bsums[2] + y[i].bsums[3])); + const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); + const __m128i q4bitsH = _mm_loadu_si128((const __m128i*)qh); - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d[0]); + const __m128i q4h_0 = _mm_slli_epi16(_mm_and_si128(q4bitsH, m2), 4); + const __m128i q4h_1 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH, 2), m2), 4); + const __m128i q4h_2 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH, 4), m2), 4); + const __m128i q4h_3 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH, 6), m2), 4); - for (int j = 0; j < QK_K/32; ++j) { - for (int l = 0; l < 16; ++l) aux16[l] = q8[l] * a[l]; - q8 += 16; a += 16; - for (int l = 0; l < 16; ++l) aux16[l] += q8[l] * a[l]; - q8 += 16; a += 16; - const float dl = d * scales[j]; - for (int l = 0; l < 8; ++l) sums[l] += dl * (aux16[l] + aux16[l+8]); - } - } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif -} -#endif + const __m128i q4_0 = _mm_or_si128(_mm_and_si128(_mm256_extractf128_si256(q4bits1, 0), m4), q4h_0); + const __m128i q4_1 = _mm_or_si128(_mm_and_si128(_mm256_extractf128_si256(q4bits1, 1), m4), q4h_1); + const __m128i q4_2 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q4bits1, 0), 4), m4), q4h_2); + const __m128i q4_3 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q4bits1, 1), 4), m4), q4h_3); -#if QK_K == 256 -void ggml_vec_dot_q5_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - assert(n % QK_K == 0); + const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); - const block_q5_K * restrict x = vx; - const block_q8_K * restrict y = vy; + __m128i q8s_0 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_0, 0)); + __m128i q8s_1 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_0, 1)); + __m128i q8s_2 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_1, 0)); + __m128i q8s_3 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_1, 1)); - const int nb = n / QK_K; + __m128i p16_0 = _mm_maddubs_epi16(q4_0, _mm256_extractf128_si256(q8_0, 0)); + __m128i p16_1 = _mm_maddubs_epi16(q4_1, _mm256_extractf128_si256(q8_0, 1)); + __m128i p16_2 = _mm_maddubs_epi16(q4_2, _mm256_extractf128_si256(q8_1, 0)); + __m128i p16_3 = _mm_maddubs_epi16(q4_3, _mm256_extractf128_si256(q8_1, 1)); - static const uint32_t kmask1 = 0x3f3f3f3f; - static const uint32_t kmask2 = 0x0f0f0f0f; - static const uint32_t kmask3 = 0x03030303; + p16_0 = _mm_sub_epi16(p16_0, q8s_0); + p16_1 = _mm_sub_epi16(p16_1, q8s_1); + p16_2 = _mm_sub_epi16(p16_2, q8s_2); + p16_3 = _mm_sub_epi16(p16_3, q8s_3); - uint32_t utmp[4]; + p16_0 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_0), p16_0); + p16_1 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_0, scale_0)), p16_1); + p16_2 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_1), p16_2); + p16_3 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_1, scale_1)), p16_3); + sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); + sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); -#ifdef __ARM_NEON + acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(MM256_SET_M128I(sumi_1, sumi_0))), acc); + } - const uint8x16_t m4b = vdupq_n_u8(0xf); - const uint8x16_t mone = vdupq_n_u8(1); - const uint8x16_t mtwo = vdupq_n_u8(2); -#if defined(__ARM_FEATURE_DOTPROD) - const int32x4_t mzero = vdupq_n_s32(0); -#endif + *s = hsum_float_8(acc); - int8x16x4_t q5bytes; +#elif defined __riscv_v_intrinsic float sumf = 0; for (int i = 0; i < nb; ++i) { - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = y[i].d * GGML_FP16_TO_FP32(x[i].dmin); + const float d_all = GGML_FP16_TO_FP32(x[i].d); - const int16x8_t q8sums = vpaddq_s16(vld1q_s16(y[i].bsums), vld1q_s16(y[i].bsums + 8)); + const uint8_t * restrict q6 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; + const int8_t * restrict scale = x[i].scales; - const uint8x8_t mins8 = vld1_u8((const uint8_t*)utmp + 8); - const int16x8_t mins = vreinterpretq_s16_u16(vmovl_u8(mins8)); - const int32x4_t prod = vaddq_s32(vmull_s16(vget_low_s16 (q8sums), vget_low_s16 (mins)), - vmull_s16(vget_high_s16(q8sums), vget_high_s16(mins))); - int32_t sumi_mins = vaddvq_s32(prod); + int32_t isum = 0; - const uint8_t * scales = (const uint8_t *)utmp; + size_t vl = 16; - const uint8_t * restrict q5 = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; + vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - uint8x16x2_t qhbits = vld1q_u8_x2(qh); + // load Q6 + vuint8mf2_t q6_0 = __riscv_vle8_v_u8mf2(q6, vl); + vuint8mf2_t q6_1 = __riscv_vle8_v_u8mf2(q6+16, vl); - uint8x16x4_t q5h; + // load qh + vuint8mf2_t qh_x = __riscv_vle8_v_u8mf2(qh, vl); - int32_t sumi = 0; + vuint8mf2_t qh0 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); + qh_x = __riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl); + vuint8mf2_t qh1 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); + qh_x = __riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl); + vuint8mf2_t qh2 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); + qh_x = __riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl); + vuint8mf2_t qh3 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); - for (int j = 0; j < QK_K/64; ++j) { + vuint8mf2_t q6h_0 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(q6_0, 0xF, vl), qh0, vl); + vuint8mf2_t q6h_1 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(q6_1, 0xF, vl), qh1, vl); + vuint8mf2_t q6h_2 = __riscv_vor_vv_u8mf2(__riscv_vsrl_vx_u8mf2(q6_0, 0x4, vl), qh2, vl); + vuint8mf2_t q6h_3 = __riscv_vor_vv_u8mf2(__riscv_vsrl_vx_u8mf2(q6_1, 0x4, vl), qh3, vl); - const uint8x16x2_t q5bits = vld1q_u8_x2(q5); q5 += 32; - const int8x16x4_t q8bytes = vld1q_s8_x4(q8); q8 += 64; + vint8mf2_t q6v_0 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_0), 32, vl); + vint8mf2_t q6v_1 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_1), 32, vl); + vint8mf2_t q6v_2 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_2), 32, vl); + vint8mf2_t q6v_3 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_3), 32, vl); - q5h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); - q5h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); - q5h.val[2] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[0]), 3); - q5h.val[3] = vshlq_n_u8(vandq_u8(mtwo, qhbits.val[1]), 3); - qhbits.val[0] = vshrq_n_u8(qhbits.val[0], 2); - qhbits.val[1] = vshrq_n_u8(qhbits.val[1], 2); + // load Q8 and take product + vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q6v_0, __riscv_vle8_v_i8mf2(q8, vl), vl); + vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q6v_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); + vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q6v_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); + vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q6v_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); - q5bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[0], m4b), q5h.val[0])); - q5bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q5bits.val[1], m4b), q5h.val[1])); - q5bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[0], 4), q5h.val[2])); - q5bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q5bits.val[1], 4), q5h.val[3])); + vint32m1_t vs_0 = __riscv_vwredsum_vs_i16m1_i32m1(p0, vzero, vl); + vint32m1_t vs_1 = __riscv_vwredsum_vs_i16m1_i32m1(p1, vzero, vl); + vint32m1_t vs_2 = __riscv_vwredsum_vs_i16m1_i32m1(p2, vzero, vl); + vint32m1_t vs_3 = __riscv_vwredsum_vs_i16m1_i32m1(p3, vzero, vl); + + isum += __riscv_vmv_x_s_i32m1_i32(vs_0) * scale[0]; + isum += __riscv_vmv_x_s_i32m1_i32(vs_1) * scale[1]; + isum += __riscv_vmv_x_s_i32m1_i32(vs_2) * scale[2]; + isum += __riscv_vmv_x_s_i32m1_i32(vs_3) * scale[3]; + + sumf += isum * d_all * y[i].d; + + } -#if defined(__ARM_FEATURE_DOTPROD) + *s = sumf; - sumi += vaddvq_s32(vdotq_s32(vdotq_s32(mzero, q5bytes.val[0], q8bytes.val[0]), q5bytes.val[1], q8bytes.val[1])) * *scales++; - sumi += vaddvq_s32(vdotq_s32(vdotq_s32(mzero, q5bytes.val[2], q8bytes.val[2]), q5bytes.val[3], q8bytes.val[3])) * *scales++; #else - const int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q5bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q5bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - sumi += vaddvq_s16(vaddq_s16(p0, p1)) * *scales++; - - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q5bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - const int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q5bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - sumi += vaddvq_s16(vaddq_s16(p2, p3)) * *scales++; -#endif + int8_t aux8[QK_K]; + int16_t aux16[8]; + float sums [8]; + int32_t aux32[8]; + memset(sums, 0, 8*sizeof(float)); + + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const uint8_t * restrict q4 = x[i].ql; + const uint8_t * restrict qh = x[i].qh; + const int8_t * restrict q8 = y[i].qs; + memset(aux32, 0, 8*sizeof(int32_t)); + int8_t * restrict a = aux8; + for (int l = 0; l < 16; ++l) { + a[l+ 0] = (int8_t)((q4[l+ 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32; + a[l+16] = (int8_t)((q4[l+16] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32; + a[l+32] = (int8_t)((q4[l+ 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32; + a[l+48] = (int8_t)((q4[l+16] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32; + } + int is = 0; + for (int j = 0; j < QK_K/16; ++j) { + int scale = x[i].scales[is++]; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; + for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; + for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; + q8 += 8; a += 8; } - - sumf += d * sumi - dmin * sumi_mins; - + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; } - + for (int l = 0; l < 8; ++l) sumf += sums[l]; *s = sumf; +#endif +} -#elif defined __AVX2__ - - const __m256i m4 = _mm256_set1_epi8(0xF); - const __m128i mzero = _mm_setzero_si128(); - const __m256i mone = _mm256_set1_epi8(1); - - __m256 acc = _mm256_setzero_ps(); - - float summs = 0.f; - - for (int i = 0; i < nb; ++i) { - - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; - -#if QK_K == 256 - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); - - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; -#else - // TODO - const float d = 0, dmin = 0; #endif - const __m256i mins_and_scales = _mm256_cvtepu8_epi16(_mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0])); +#if defined (__AVX2__) || defined (__ARM_NEON) +static const int8_t keven_signs_q2xs[1024] = { + 1, 1, 1, 1, 1, 1, 1, 1, -1, 1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, 1, + 1, 1, -1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, 1, 1, 1, -1, -1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, -1, + 1, 1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, -1, + 1, 1, -1, -1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, 1, + 1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, 1, 1, 1, -1, 1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, -1, + 1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, -1, 1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, 1, + 1, 1, 1, -1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, 1, + 1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, 1, 1, -1, -1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, -1, + 1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, -1, 1, -1, + 1, 1, -1, 1, 1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, 1, + 1, 1, 1, -1, 1, -1, 1, 1, -1, 1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, 1, + 1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, -1, + 1, 1, 1, 1, -1, -1, 1, 1, -1, 1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, 1, + 1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, 1, 1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, -1, + 1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, 1, 1, 1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, -1, + 1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, 1, + 1, 1, 1, 1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, -1, 1, 1, -1, 1, 1, 1, 1, -1, 1, -1, -1, 1, 1, 1, 1, -1, -1, + 1, 1, -1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, 1, -1, -1, 1, -1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, 1, + 1, 1, 1, -1, 1, 1, -1, 1, -1, 1, 1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, 1, + 1, 1, -1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, 1, 1, -1, -1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, -1, + 1, 1, 1, 1, -1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, -1, 1, -1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, 1, + 1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, 1, 1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, -1, + 1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, 1, -1, 1, 1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, -1, + 1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, 1, -1, -1, 1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, 1, + 1, 1, 1, 1, 1, -1, -1, 1, -1, 1, 1, 1, 1, -1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, -1, 1, + 1, 1, -1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, 1, -1, -1, 1, -1, -1, -1, 1, 1, -1, -1, -1, + 1, 1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, 1, -1, -1, 1, 1, -1, 1, -1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, -1, + 1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, -1, 1, -1, -1, -1, 1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, 1, + 1, 1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, -1, -1, -1, 1, 1, -1, 1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, -1, -1, + 1, 1, -1, 1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, -1, 1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, 1, + 1, 1, 1, -1, -1, -1, -1, 1, -1, 1, 1, -1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, 1, + 1, 1, -1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, 1, 1, -1, -1, -1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, -1, +}; +#endif - const __m256i q8sums = _mm256_loadu_si256((const __m256i*)y[i].bsums); - const __m128i q8s = _mm_hadd_epi16(_mm256_extracti128_si256(q8sums, 0), _mm256_extracti128_si256(q8sums, 1)); - const __m128i prod = _mm_madd_epi16(_mm256_extracti128_si256(mins_and_scales, 1), q8s); - const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); - summs += dmin * _mm_extract_epi32(hsum, 0); +void ggml_vec_dot_iq2_xxs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const __m128i sc128 = _mm256_extracti128_si256(mins_and_scales, 0); - const __m256i scales = MM256_SET_M128I(sc128, sc128); + const block_iq2_xxs * restrict x = vx; + const block_q8_K * restrict y = vy; - const __m256i hbits = _mm256_loadu_si256((const __m256i*)x[i].qh); - __m256i hmask = mone; + const int nb = n / QK_K; - __m256i sumi = _mm256_setzero_si256(); +#if defined(__ARM_NEON) - int bit = 0; + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - for (int j = 0; j < QK_K/64; ++j) { + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; - const __m256i scale_0 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+0)); - const __m256i scale_1 = _mm256_shuffle_epi8(scales, get_scale_shuffle_k4(2*j+1)); + ggml_int8x16x4_t q2u; + ggml_int8x16x4_t q2s; + ggml_int8x16x4_t q8b; - const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); q5 += 32; + float sumf = 0; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + float sumf1 = 0, sumf2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; + q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 0])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 1]))); + q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 2])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 3]))); + q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[ 8])), vld1_s8((const void *)(iq2xxs_grid + aux8[ 9]))); + q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xxs_grid + aux8[10])), vld1_s8((const void *)(iq2xxs_grid + aux8[11]))); + q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 7) & 127)))); + q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 21) & 127)))); + q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[3] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[3] >> 7) & 127)))); + q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[3] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[3] >> 21) & 127)))); + q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]); + q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]); + q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]); + q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]); + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]), q2u.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]), q2u.val[3], q8b.val[3]); + sumf1 += vaddvq_s32(p1) * (0.5f + (aux32[1] >> 28)); + sumf2 += vaddvq_s32(p2) * (0.5f + (aux32[3] >> 28)); + } + sumf += d*(sumf1 + sumf2); + } + *s = 0.25f * sumf; - const __m256i q5l_0 = _mm256_and_si256(q5bits, m4); - const __m256i q5h_0 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); - const __m256i q5_0 = _mm256_add_epi8(q5l_0, q5h_0); - hmask = _mm256_slli_epi16(hmask, 1); +#elif defined(__AVX2__) - const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4); - const __m256i q5h_1 = _mm256_slli_epi16(_mm256_srli_epi16(_mm256_and_si256(hbits, hmask), bit++), 4); - const __m256i q5_1 = _mm256_add_epi8(q5l_1, q5h_1); - hmask = _mm256_slli_epi16(hmask, 1); + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + uint32_t aux32[4]; + const uint8_t * aux8 = (const uint8_t *)aux32; - __m256i p16_0 = _mm256_maddubs_epi16(q5_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q5_1, q8_1); + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + memcpy(aux32, q2, 4*sizeof(uint32_t)); q2 += 8; + const __m256i q2_1 = _mm256_set_epi64x(iq2xxs_grid[aux8[ 3]], iq2xxs_grid[aux8[ 2]], iq2xxs_grid[aux8[1]], iq2xxs_grid[aux8[0]]); + const __m256i q2_2 = _mm256_set_epi64x(iq2xxs_grid[aux8[11]], iq2xxs_grid[aux8[10]], iq2xxs_grid[aux8[9]], iq2xxs_grid[aux8[8]]); + const __m256i s2_1 = _mm256_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], + signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m256i s2_2 = _mm256_set_epi64x(signs64[(aux32[3] >> 21) & 127], signs64[(aux32[3] >> 14) & 127], + signs64[(aux32[3] >> 7) & 127], signs64[(aux32[3] >> 0) & 127]); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2); + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const uint16_t ls1 = aux32[1] >> 28; + const uint16_t ls2 = aux32[3] >> 28; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); - p16_0 = _mm256_madd_epi16(scale_0, p16_0); - p16_1 = _mm256_madd_epi16(scale_1, p16_1); +#else - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); + uint32_t aux32[2]; + const uint8_t * aux8 = (const uint8_t *)aux32; + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + memcpy(aux32, q2, 2*sizeof(uint32_t)); + q2 += 4; + const uint32_t ls = 2*(aux32[1] >> 28) + 1; + int32_t sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xxs_grid + aux8[l]); + const uint8_t signs = ksigns_iq2xs[(aux32[1] >> 7*l) & 127]; + for (int j = 0; j < 8; ++j) { + sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += sumi * ls; } + sumf += d * bsum; + } + *s = 0.125f * sumf; +#endif +} - __m256 vd = _mm256_set1_ps(d); - acc = _mm256_fmadd_ps(vd, _mm256_cvtepi32_ps(sumi), acc); +void ggml_vec_dot_iq2_xs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - } + const block_iq2_xs * restrict x = vx; + const block_q8_K * restrict y = vy; - *s = hsum_float_8(acc) + summs; + const int nb = n / QK_K; -#elif defined __AVX__ +#if defined(__ARM_NEON) - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i mzero = _mm_setzero_si128(); - const __m128i mone = _mm_set1_epi8(1); - const __m128i m2 = _mm_set1_epi8(2); + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; - __m256 acc = _mm256_setzero_ps(); + ggml_int8x16x4_t q2u; + ggml_int8x16x4_t q2s; + ggml_int8x16x4_t q8b; - float summs = 0.f; + int32x4x4_t scales32; + float sumf = 0; for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; + const uint8x8_t scales8 = vld1_u8(x[i].scales); + const uint8x8_t scales_l = vand_u8(scales8, vdup_n_u8(0xf)); + const uint8x8_t scales_h = vshr_n_u8(scales8, 4); + uint8x16_t scales = vcombine_u8(vzip1_u8(scales_l, scales_h), vzip2_u8(scales_l, scales_h)); + scales = vaddq_u8(vshlq_n_u8(scales, 1), vdupq_n_u8(1)); + const uint16x8_t scales1 = vmovl_u8(vget_low_u8(scales)); + const uint16x8_t scales2 = vmovl_u8(vget_high_u8(scales)); + scales32.val[0] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales1))); + scales32.val[1] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales1))); + scales32.val[2] = vreinterpretq_s32_u32(vmovl_u16(vget_low_u16(scales2))); + scales32.val[3] = vreinterpretq_s32_u32(vmovl_u16(vget_high_u16(scales2))); + int32x4_t sumi = vdupq_n_s32(0); + for (int ib64 = 0; ib64 < QK_K/64; ++ib64) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + q2u.val[0] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[0] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[1] & 511)))); + q2u.val[1] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[2] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[3] & 511)))); + q2u.val[2] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[4] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[5] & 511)))); + q2u.val[3] = vcombine_s8(vld1_s8((const void *)(iq2xs_grid + (q2[6] & 511))), vld1_s8((const void *)(iq2xs_grid + (q2[7] & 511)))); + q2s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[0] >> 9))), vld1_s8((const void *)(signs64 + (q2[1] >> 9)))); + q2s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[2] >> 9))), vld1_s8((const void *)(signs64 + (q2[3] >> 9)))); + q2s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[4] >> 9))), vld1_s8((const void *)(signs64 + (q2[5] >> 9)))); + q2s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + (q2[6] >> 9))), vld1_s8((const void *)(signs64 + (q2[7] >> 9)))); + q2u.val[0] = vmulq_s8(q2u.val[0], q2s.val[0]); + q2u.val[1] = vmulq_s8(q2u.val[1], q2s.val[1]); + q2u.val[2] = vmulq_s8(q2u.val[2], q2s.val[2]); + q2u.val[3] = vmulq_s8(q2u.val[3], q2s.val[3]); + const int32x4_t p1 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[0], q8b.val[0]); + const int32x4_t p2 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[1], q8b.val[1]); + const int32x4_t p3 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[2], q8b.val[2]); + const int32x4_t p4 = ggml_vdotq_s32(vdupq_n_s32(0), q2u.val[3], q8b.val[3]); + const int32x4_t p = vpaddq_s32(vpaddq_s32(p1, p2), vpaddq_s32(p3, p4)); + sumi = vmlaq_s32(sumi, p, scales32.val[ib64]); + q2 += 8; + } + sumf += d*vaddvq_s32(sumi); + } + *s = 0.125f * sumf; - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const float dmin = -y[i].d * GGML_FP16_TO_FP32(x[i].dmin); +#elif defined(__AVX2__) - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; + const __m256i mone = _mm256_set1_epi8(1); + static const char block_sign_shuffle_mask_1[32] = { + 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, + 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x04, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, 0x06, + }; + static const char block_sign_shuffle_mask_2[32] = { + 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x08, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, + 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, 0x0e, + }; + static const uint8_t bit_selector_mask_bytes[32] = { + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; + const __m256i bit_selector_mask = _mm256_loadu_si256((const __m256i*)bit_selector_mask_bytes); + const __m256i block_sign_shuffle_1 = _mm256_loadu_si256((const __m256i*)block_sign_shuffle_mask_1); + const __m256i block_sign_shuffle_2 = _mm256_loadu_si256((const __m256i*)block_sign_shuffle_mask_2); - const __m128i utmps = _mm_set_epi32(utmp[3], utmp[2], utmp[1], utmp[0]); - const __m128i scales = _mm_cvtepu8_epi16(utmps); - const __m128i mins = _mm_cvtepu8_epi16(_mm_unpackhi_epi64(utmps, utmps)); +#if QK_K == 64 + static const uint8_t k_bit_helper[16] = { + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + }; + const __m128i bit_helper = _mm_loadu_si128((const __m128i*)k_bit_helper); + const __m128i m511 = _mm_set1_epi16(511); + typedef union { + __m128i vec_index; + uint16_t index[8]; + } index_t; + + index_t idx; + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const __m128i q2_data = _mm_loadu_si128((const __m128i*)x[i].qs); + idx.vec_index = _mm_and_si128(q2_data, m511); - const __m128i q8sums_0 = _mm_loadu_si128((const __m128i*)&y[i].bsums[0]); - const __m128i q8sums_1 = _mm_loadu_si128((const __m128i*)&y[i].bsums[8]); - const __m128i q8s = _mm_hadd_epi16(q8sums_0, q8sums_1); - const __m128i prod = _mm_madd_epi16(mins, q8s); - const __m128i hsum = _mm_hadd_epi32(_mm_hadd_epi32(prod, mzero), mzero); - summs += dmin * _mm_extract_epi32(hsum, 0); + const __m128i partial_sign_bits = _mm_srli_epi16(q2_data, 9); + const __m128i partial_sign_bits_upper = _mm_srli_epi16(q2_data, 13); + const __m128i partial_sign_bits_for_counting = _mm_xor_si128(partial_sign_bits, partial_sign_bits_upper); - const __m128i hbits_0 = _mm_loadu_si128((const __m128i*)&x[i].qh[0]); - const __m128i hbits_1 = _mm_loadu_si128((const __m128i*)&x[i].qh[16]); - __m128i hmask = mone; + const __m128i odd_bits = _mm_shuffle_epi8(bit_helper, partial_sign_bits_for_counting); + const __m128i full_sign_bits = _mm_or_si128(partial_sign_bits, odd_bits); + const __m256i full_signs = MM256_SET_M128I(full_sign_bits, full_sign_bits); - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)y[i].qs); + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)(y[i].qs+32)); - int bit = 0; + const __m256i q2_1 = _mm256_set_epi64x(iq2xs_grid[idx.index[3]], iq2xs_grid[idx.index[2]], + iq2xs_grid[idx.index[1]], iq2xs_grid[idx.index[0]]); + const __m256i q2_2 = _mm256_set_epi64x(iq2xs_grid[idx.index[7]], iq2xs_grid[idx.index[6]], + iq2xs_grid[idx.index[5]], iq2xs_grid[idx.index[4]]); - __m128i shuffle = _mm_set1_epi16(0x0100); - for (int j = 0; j < QK_K/64; ++j) { + __m256i signs; + signs = _mm256_shuffle_epi8(full_signs, block_sign_shuffle_1); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, _mm256_or_si256(signs, mone)); - const __m128i scale_0 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); - const __m128i scale_1 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi16(shuffle, m2); + signs = _mm256_shuffle_epi8(full_signs, block_sign_shuffle_2); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, _mm256_or_si256(signs, mone)); - const __m128i q5bits_0 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; - const __m128i q5bits_1 = _mm_loadu_si128((const __m128i*)q5); q5 += 16; + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); - __m128i q5l_0 = _mm_and_si128(q5bits_0, m4); - __m128i q5l_1 = _mm_and_si128(q5bits_1, m4); - __m128i q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); - __m128i q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); - __m128i q5_0 = _mm_add_epi8(q5l_0, q5h_0); - __m128i q5_1 = _mm_add_epi8(q5l_1, q5h_1); - hmask = _mm_slli_epi16(hmask, 1); + const __m256i sc1 = MM256_SET_M128I(_mm_set1_epi16(2*(x[i].scales[0] >> 4)+1), _mm_set1_epi16(2*(x[i].scales[0] & 0xf)+1)); + const __m256i sc2 = MM256_SET_M128I(_mm_set1_epi16(2*(x[i].scales[1] >> 4)+1), _mm_set1_epi16(2*(x[i].scales[1] & 0xf)+1)); - __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16_0 = _mm_maddubs_epi16(q5_0, q8_0); - __m128i p16_1 = _mm_maddubs_epi16(q5_1, q8_1); - p16_0 = _mm_madd_epi16(scale_0, p16_0); - p16_1 = _mm_madd_epi16(scale_0, p16_1); + const __m256i sum = _mm256_add_epi32(_mm256_madd_epi16(sc1, dot1), _mm256_madd_epi16(sc2, dot2)); - q5l_0 = _mm_and_si128(_mm_srli_epi16(q5bits_0, 4), m4); - q5l_1 = _mm_and_si128(_mm_srli_epi16(q5bits_1, 4), m4); - q5h_0 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_0, hmask), bit), 4); - q5h_1 = _mm_slli_epi16(_mm_srli_epi16(_mm_and_si128(hbits_1, hmask), bit++), 4); - q5_0 = _mm_add_epi8(q5l_0, q5h_0); - q5_1 = _mm_add_epi8(q5l_1, q5h_1); - hmask = _mm_slli_epi16(hmask, 1); + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(sum), accumf); - q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - __m128i p16_2 = _mm_maddubs_epi16(q5_0, q8_0); - __m128i p16_3 = _mm_maddubs_epi16(q5_1, q8_1); - p16_2 = _mm_madd_epi16(scale_1, p16_2); - p16_3 = _mm_madd_epi16(scale_1, p16_3); + } - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); + *s = 0.125f * hsum_float_8(accumf); +#else - } + static const uint8_t k_bit_helper[32] = { + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + 0x00, 0x80, 0x80, 0x00, 0x80, 0x00, 0x00, 0x80, 0x80, 0x00, 0x00, 0x80, 0x00, 0x80, 0x80, 0x00, + }; + const __m256i bit_helper = _mm256_loadu_si256((const __m256i*)k_bit_helper); + const __m256i m511 = _mm256_set1_epi16(511); + const __m128i m4 = _mm_set1_epi8(0xf); + const __m128i m1 = _mm_set1_epi8(1); - __m256 vd = _mm256_set1_ps(d); - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(vd, _mm256_cvtepi32_ps(sumi)), acc); + uint64_t aux64; - } + // somewhat hacky, but gives a significant boost in performance + __m256i aux_gindex; + const uint16_t * gindex = (const uint16_t *)&aux_gindex; - *s = hsum_float_8(acc) + summs; + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const int8_t * restrict q8 = y[i].qs; -#elif defined __riscv_v_intrinsic + memcpy(&aux64, x[i].scales, 8); + __m128i stmp = _mm_set1_epi64x(aux64); + stmp = _mm_unpacklo_epi8(_mm_and_si128(stmp, m4), _mm_and_si128(_mm_srli_epi16(stmp, 4), m4)); + const __m128i scales = _mm_add_epi8(_mm_slli_epi16(stmp, 1), m1); - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 4) { - float sumf = 0; - float sums = 0.0; + const __m256i q2_data = _mm256_loadu_si256((const __m256i*)q2); q2 += 16; + aux_gindex = _mm256_and_si256(q2_data, m511); - size_t vl; + const __m256i partial_sign_bits = _mm256_srli_epi16(q2_data, 9); + const __m256i partial_sign_bits_upper = _mm256_srli_epi16(q2_data, 13); + const __m256i partial_sign_bits_for_counting = _mm256_xor_si256(partial_sign_bits, partial_sign_bits_upper); - for (int i = 0; i < nb; ++i) { + const __m256i odd_bits = _mm256_shuffle_epi8(bit_helper, partial_sign_bits_for_counting); + const __m256i full_sign_bits = _mm256_or_si256(partial_sign_bits, odd_bits); - vl = 8; + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_3 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_4 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; - const uint8_t * restrict q5 = x[i].qs; - const uint8_t * restrict hm = x[i].qh; - const int8_t * restrict q8 = y[i].qs; + const __m256i q2_1 = _mm256_set_epi64x(iq2xs_grid[gindex[ 3]], iq2xs_grid[gindex[ 2]], + iq2xs_grid[gindex[ 1]], iq2xs_grid[gindex[ 0]]); + const __m256i q2_2 = _mm256_set_epi64x(iq2xs_grid[gindex[ 7]], iq2xs_grid[gindex[ 6]], + iq2xs_grid[gindex[ 5]], iq2xs_grid[gindex[ 4]]); + const __m256i q2_3 = _mm256_set_epi64x(iq2xs_grid[gindex[11]], iq2xs_grid[gindex[10]], + iq2xs_grid[gindex[ 9]], iq2xs_grid[gindex[ 8]]); + const __m256i q2_4 = _mm256_set_epi64x(iq2xs_grid[gindex[15]], iq2xs_grid[gindex[14]], + iq2xs_grid[gindex[13]], iq2xs_grid[gindex[12]]); - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; + const __m128i full_signs_l = _mm256_castsi256_si128(full_sign_bits); + const __m128i full_signs_h = _mm256_extractf128_si256(full_sign_bits, 1); + const __m256i full_signs_1 = MM256_SET_M128I(full_signs_l, full_signs_l); + const __m256i full_signs_2 = MM256_SET_M128I(full_signs_h, full_signs_h); - vint16mf2_t q8sums_0 = __riscv_vlse16_v_i16mf2(y[i].bsums, 4, vl); - vint16mf2_t q8sums_1 = __riscv_vlse16_v_i16mf2(y[i].bsums+1, 4, vl); - vint16mf2_t q8sums = __riscv_vadd_vv_i16mf2(q8sums_0, q8sums_1, vl); + __m256i signs; + signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_1); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, _mm256_or_si256(signs, mone)); - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; + signs = _mm256_shuffle_epi8(full_signs_1, block_sign_shuffle_2); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, _mm256_or_si256(signs, mone)); - vuint8mf4_t mins8 = __riscv_vle8_v_u8mf4(mins, vl); - vint16mf2_t v_mins = __riscv_vreinterpret_v_u16mf2_i16mf2(__riscv_vzext_vf2_u16mf2(mins8, vl)); - vint32m1_t prod = __riscv_vwmul_vv_i32m1(q8sums, v_mins, vl); + signs = _mm256_shuffle_epi8(full_signs_2, block_sign_shuffle_1); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_3 = _mm256_sign_epi8(q8_3, _mm256_or_si256(signs, mone)); - vint32m1_t sumi = __riscv_vredsum_vs_i32m1_i32m1(prod, __riscv_vmv_v_x_i32m1(0, 1), vl); - sumf -= dmin * __riscv_vmv_x_s_i32m1_i32(sumi); + signs = _mm256_shuffle_epi8(full_signs_2, block_sign_shuffle_2); + signs = _mm256_cmpeq_epi8(_mm256_and_si256(signs, bit_selector_mask), bit_selector_mask); + const __m256i q8s_4 = _mm256_sign_epi8(q8_4, _mm256_or_si256(signs, mone)); - vl = 32; - int32_t aux32 = 0; - int is = 0; + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const __m256i dot3 = _mm256_maddubs_epi16(q2_3, q8s_3); + const __m256i dot4 = _mm256_maddubs_epi16(q2_4, q8s_4); - uint8_t m = 1; - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); - vuint8m1_t vqh = __riscv_vle8_v_u8m1(hm, vl); + const __m256i sc1 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+0))); + const __m256i sc2 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+1))); + const __m256i sc3 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+2))); + const __m256i sc4 = _mm256_cvtepi8_epi16(_mm_shuffle_epi8(scales, get_scale_shuffle(ib32+3))); - for (int j = 0; j < QK_K/64; ++j) { - // load Q5 and Q8 - vuint8m1_t q5_x = __riscv_vle8_v_u8m1(q5, vl); - vint8m1_t q8_y1 = __riscv_vle8_v_i8m1(q8, vl); - vint8m1_t q8_y2 = __riscv_vle8_v_i8m1(q8+32, vl); + sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot1, sc1)); + sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot2, sc2)); + sumi1 = _mm256_add_epi32(sumi1, _mm256_madd_epi16(dot3, sc3)); + sumi2 = _mm256_add_epi32(sumi2, _mm256_madd_epi16(dot4, sc4)); + } - // compute mask for addition - vint8m1_t q5_a = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vand_vx_u8m1(q5_x, 0x0F, vl)); - vuint8m1_t qh_m1 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_1 = __riscv_vmsne_vx_u8m1_b8(qh_m1, 0, vl); - vint8m1_t q5_m1 = __riscv_vadd_vx_i8m1_m(vmask_1, q5_a, 16, vl); - m <<= 1; + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); - vint8m1_t q5_l = __riscv_vreinterpret_v_u8m1_i8m1(__riscv_vsrl_vx_u8m1(q5_x, 0x04, vl)); - vuint8m1_t qh_m2 = __riscv_vand_vx_u8m1(vqh, m, vl); - vbool8_t vmask_2 = __riscv_vmsne_vx_u8m1_b8(qh_m2, 0, vl); - vint8m1_t q5_m2 = __riscv_vadd_vx_i8m1_m(vmask_2, q5_l, 16, vl); - m <<= 1; + } - vint16m2_t v0 = __riscv_vwmul_vv_i16m2(q5_m1, q8_y1, vl); - vint16m2_t v1 = __riscv_vwmul_vv_i16m2(q5_m2, q8_y2, vl); + *s = 0.125f * hsum_float_8(accumf); +#endif - vint32m4_t vs1 = __riscv_vwmul_vx_i32m4(v0, scales[is++], vl); - vint32m4_t vs2 = __riscv_vwmul_vx_i32m4(v1, scales[is++], vl); +#else - vint32m1_t vacc1 = __riscv_vredsum_vs_i32m4_i32m1(vs1, vzero, vl); - vint32m1_t vacc2 = __riscv_vredsum_vs_i32m4_i32m1(vs2, vzero, vl); + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint16_t * restrict q2 = x[i].qs; + const uint8_t * restrict sc = x[i].scales; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + const uint16_t ls1 = 2*(sc[ib32] & 0xf) + 1; + const uint16_t ls2 = 2*(sc[ib32] >> 4) + 1; + int32_t sumi = 0; + for (int l = 0; l < 2; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511)); + const uint8_t signs = ksigns_iq2xs[q2[l] >> 9]; + for (int j = 0; j < 8; ++j) { + sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += sumi * ls1; + sumi = 0; + for (int l = 2; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2xs_grid + (q2[l] & 511)); + const uint8_t signs = ksigns_iq2xs[q2[l] >> 9]; + for (int j = 0; j < 8; ++j) { + sumi += grid[j] * q8[j] * (signs & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += sumi * ls2; + q2 += 4; + } + sumf += d * bsum; + } + *s = 0.125f * sumf; +#endif +} - aux32 += __riscv_vmv_x_s_i32m1_i32(vacc1) + __riscv_vmv_x_s_i32m1_i32(vacc2); - q5 += 32; q8 += 64; +void ggml_vec_dot_iq2_s_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - } + const block_iq2_s * restrict x = vx; + const block_q8_K * restrict y = vy; - vfloat32m1_t vaux = __riscv_vfmul_vf_f32m1(__riscv_vfmv_v_f_f32m1(aux32, 1), d, 1); - sums += __riscv_vfmv_f_s_f32m1_f32(vaux); + const int nb = n / QK_K; - } +#if defined(__ARM_NEON) - *s = sumf+sums; + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; -#else + static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; - const uint8_t * scales = (const uint8_t*)&utmp[0]; - const uint8_t * mins = (const uint8_t*)&utmp[2]; + const ggml_uint8x16x2_t mask1 = ggml_vld1q_u8_x2(k_mask1); + const uint8x16_t mask2 = vld1q_u8(k_mask2); + const uint8x16_t m1 = vdupq_n_u8(1); + const int32x4_t vzero = vdupq_n_s32(0); - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); + uint8x16x2_t vs; + ggml_int8x16x4_t q2s; + ggml_int8x16x4_t q8b; float sumf = 0; for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; - const uint8_t * restrict hm = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - uint8_t m = 1; - for (int j = 0; j < QK_K/64; ++j) { - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] & 0xF); - for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); - a += 32; m <<= 1; - for (int l = 0; l < 32; ++l) a[l] = (int8_t)(q4[l] >> 4); - for (int l = 0; l < 32; ++l) a[l] += (hm[l] & m ? 16 : 0); - a += 32; m <<= 1; - q4 += 32; - } - memcpy(utmp, x[i].scales, 12); - utmp[3] = ((utmp[2] >> 4) & kmask2) | (((utmp[1] >> 6) & kmask3) << 4); - const uint32_t uaux = utmp[1] & kmask1; - utmp[1] = (utmp[2] & kmask2) | (((utmp[0] >> 6) & kmask3) << 4); - utmp[2] = uaux; - utmp[0] &= kmask1; - int sumi = 0; - for (int j = 0; j < QK_K/16; ++j) sumi += y[i].bsums[j] * mins[j/2]; - a = aux8; - int is = 0; - for (int j = 0; j < QK_K/32; ++j) { - int32_t scale = scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - } const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; - const float dmin = GGML_FP16_TO_FP32(x[i].dmin) * y[i].d; - sumf -= dmin * sumi; - } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif -} -#else + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); + const int8_t * restrict q8 = y[i].qs; -void ggml_vec_dot_q5_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - assert(n % QK_K == 0); + int sumi1 = 0, sumi2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + q2s.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[0] | ((qh[ib32+0] << 8) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[1] | ((qh[ib32+0] << 6) & 0x300))))); + q2s.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[2] | ((qh[ib32+0] << 4) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[3] | ((qh[ib32+0] << 2) & 0x300))))); + q2s.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[4] | ((qh[ib32+1] << 8) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[5] | ((qh[ib32+1] << 6) & 0x300))))); + q2s.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq2s_grid + (qs[6] | ((qh[ib32+1] << 4) & 0x300)))), + vld1_s8((const int8_t *)(iq2s_grid + (qs[7] | ((qh[ib32+1] << 2) & 0x300))))); + qs += 8; - const block_q5_K * restrict x = vx; - const block_q8_K * restrict y = vy; + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | ((uint32_t) signs[1] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vceqq_u8(vs.val[0], mask2); + vs.val[1] = vceqq_u8(vs.val[1], mask2); - const int nb = n / QK_K; + q2s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[0], m1)), q2s.val[0]); + q2s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[1]); -#ifdef __ARM_NEON + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | ((uint32_t) signs[3] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vceqq_u8(vs.val[0], mask2); + vs.val[1] = vceqq_u8(vs.val[1], mask2); - const uint8x16_t m4b = vdupq_n_u8(0xf); - const uint8x16_t mh = vdupq_n_u8(16); -#if defined(__ARM_FEATURE_DOTPROD) - const int32x4_t mzero = vdupq_n_s32(0); -#endif + signs += 4; - int8x16x4_t q5bytes; - uint8x16x4_t q5h; + q2s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[0], m1)), q2s.val[2]); + q2s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vorrq_u8(vs.val[1], m1)), q2s.val[3]); - float sumf = 0; + const int32x4_t p1 = ggml_vdotq_s32(vzero, q2s.val[0], q8b.val[0]); + const int32x4_t p2 = ggml_vdotq_s32(vzero, q2s.val[1], q8b.val[1]); + const int32x4_t p3 = ggml_vdotq_s32(vzero, q2s.val[2], q8b.val[2]); + const int32x4_t p4 = ggml_vdotq_s32(vzero, q2s.val[3], q8b.val[3]); - for (int i = 0; i < nb; ++i) { + sumi1 += vaddvq_s32(p1) * (1 + 2*(x[i].scales[ib32+0] & 0xf)); + sumi2 += vaddvq_s32(p2) * (1 + 2*(x[i].scales[ib32+0] >> 4)); + sumi1 += vaddvq_s32(p3) * (1 + 2*(x[i].scales[ib32+1] & 0xf)); + sumi2 += vaddvq_s32(p4) * (1 + 2*(x[i].scales[ib32+1] >> 4)); + } + sumf += d*(sumi1 + sumi2); + } - const float d = y[i].d * (float)x[i].d; - const int8_t * sc = x[i].scales; + *s = 0.125f * sumf; - const uint8_t * restrict q5 = x[i].qs; +#elif defined(__AVX2__) + + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; + + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; + + const __m128i m4 = _mm_set1_epi8(0xf); + const __m128i m1 = _mm_set1_epi8(1); + + const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1); + const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2); + + uint64_t aux64; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)(x[i].qs + QK_K/8); const int8_t * restrict q8 = y[i].qs; - const uint8x8_t qhbits = vld1_u8(qh); + memcpy(&aux64, x[i].scales, 8); + const __m128i scales8 = _mm_add_epi8(_mm_slli_epi16(_mm_and_si128(_mm_set_epi64x(aux64 >> 4, aux64), m4), 1), m1); + const __m256i scales16 = _mm256_cvtepi8_epi16(scales8); // 0 2 4 6 8 10 12 14 1 3 5 7 9 11 13 15 - const uint8x16x2_t q5bits = vld1q_u8_x2(q5); - const int8x16x4_t q8bytes = vld1q_s8_x4(q8); + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q2_1 = _mm256_set_epi64x(iq2s_grid[qs[3] | ((qh[ib32+0] << 2) & 0x300)], + iq2s_grid[qs[2] | ((qh[ib32+0] << 4) & 0x300)], + iq2s_grid[qs[1] | ((qh[ib32+0] << 6) & 0x300)], + iq2s_grid[qs[0] | ((qh[ib32+0] << 8) & 0x300)]); + const __m256i q2_2 = _mm256_set_epi64x(iq2s_grid[qs[7] | ((qh[ib32+1] << 2) & 0x300)], + iq2s_grid[qs[6] | ((qh[ib32+1] << 4) & 0x300)], + iq2s_grid[qs[5] | ((qh[ib32+1] << 6) & 0x300)], + iq2s_grid[qs[4] | ((qh[ib32+1] << 8) & 0x300)]); + qs += 8; - const uint8x16_t htmp = vcombine_u8(qhbits, vshr_n_u8(qhbits, 1)); - q5h.val[0] = vbicq_u8(mh, vshlq_n_u8(htmp, 4)); - q5h.val[1] = vbicq_u8(mh, vshlq_n_u8(htmp, 2)); - q5h.val[2] = vbicq_u8(mh, htmp); - q5h.val[3] = vbicq_u8(mh, vshrq_n_u8(htmp, 2)); + __m256i aux256 = _mm256_set1_epi32(signs[0] | ((uint32_t) signs[1] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_1 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_1 = _mm256_sub_epi8(_mm256_xor_si256(s2_1, q8_1), s2_1); - q5bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q5bits.val[0], m4b)), vreinterpretq_s8_u8(q5h.val[0])); - q5bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vandq_u8(q5bits.val[1], m4b)), vreinterpretq_s8_u8(q5h.val[1])); - q5bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vshrq_n_u8(q5bits.val[0], 4)), vreinterpretq_s8_u8(q5h.val[2])); - q5bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vshrq_n_u8(q5bits.val[1], 4)), vreinterpretq_s8_u8(q5h.val[3])); + aux256 = _mm256_set1_epi32(signs[2] | ((uint32_t) signs[3] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_2 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_2 = _mm256_sub_epi8(_mm256_xor_si256(s2_2, q8_2), s2_2); -#if defined(__ARM_FEATURE_DOTPROD) + signs += 4; - int32_t sumi1 = sc[0] * vaddvq_s32(vdotq_s32(mzero, q5bytes.val[0], q8bytes.val[0])); - int32_t sumi2 = sc[1] * vaddvq_s32(vdotq_s32(mzero, q5bytes.val[1], q8bytes.val[1])); - int32_t sumi3 = sc[2] * vaddvq_s32(vdotq_s32(mzero, q5bytes.val[2], q8bytes.val[2])); - int32_t sumi4 = sc[3] * vaddvq_s32(vdotq_s32(mzero, q5bytes.val[3], q8bytes.val[3])); + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); // blocks 2*ib32+0, 2*ib32+1 + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); // blocks 2*ib32+2, 2*ib32+3 - sumf += d * (sumi1 + sumi2 + sumi3 + sumi4); + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_shuffle_epi8(scales16, get_scale_shuffle_k4(ib32+0))); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_shuffle_epi8(scales16, get_scale_shuffle_k4(ib32+1))); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.125f * hsum_float_8(accumf); #else - const int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q5bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - const int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q5bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - int32_t sumi = sc[0] * vaddvq_s16(p0) + sc[1] * vaddvq_s16(p1); + float sumf = 0; + for (int i = 0; i < nb; i++) { + + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint8_t * signs = qs + QK_K/8; + + int bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + int ls1 = 1 + 2*(x[i].scales[ib32] & 0xf); + int ls2 = 1 + 2*(x[i].scales[ib32] >> 4); + int sumi1 = 0, sumi2 = 0; + for (int l = 0; l < 2; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); + for (int j = 0; j < 8; ++j) { + sumi1 += q8[j] * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + for (int l = 2; l < 4; ++l) { + const uint8_t * grid = (const uint8_t *)(iq2s_grid + (qs[l] | (qh[ib32] << (8-2*l) & 0x300))); + for (int j = 0; j < 8; ++j) { + sumi2 += q8[j] * grid[j] * (signs[l] & kmask_iq2xs[j] ? -1 : 1); + } + q8 += 8; + } + bsum += ls1 * sumi1 + ls2 * sumi2; + qs += 4; + signs += 4; + } + + sumf += d * bsum; + } - const int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q5bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - const int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q5bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q5bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - sumi += sc[2] * vaddvq_s16(p2) + sc[3] * vaddvq_s16(p3); + *s = 0.125f * sumf; - sumf += d*sumi; #endif - } +} - *s = sumf; +void ggml_vec_dot_iq3_xxs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); -#elif defined __AVX2__ + const block_iq3_xxs * restrict x = vx; + const block_q8_K * restrict y = vy; - const __m256i m4 = _mm256_set1_epi8(0xF); - const __m256i mone = _mm256_set1_epi8(1); + const int nb = n / QK_K; - __m256 acc = _mm256_setzero_ps(); +#if defined(__ARM_NEON) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[2]; + + ggml_int8x16x4_t q3s; + ggml_int8x16x4_t q8b; + float sumf = 0; for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + float sumf1 = 0, sumf2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + memcpy(aux32, gas, 2*sizeof(uint32_t)); gas += 2*sizeof(uint32_t); + const uint32x4_t aux32x4_0 = ggml_vld1q_u32(iq3xxs_grid[q3[ 0]], iq3xxs_grid[q3[ 1]], iq3xxs_grid[q3[ 2]], iq3xxs_grid[q3[ 3]]); + const uint32x4_t aux32x4_1 = ggml_vld1q_u32(iq3xxs_grid[q3[ 4]], iq3xxs_grid[q3[ 5]], iq3xxs_grid[q3[ 6]], iq3xxs_grid[q3[ 7]]); + const uint32x4_t aux32x4_2 = ggml_vld1q_u32(iq3xxs_grid[q3[ 8]], iq3xxs_grid[q3[ 9]], iq3xxs_grid[q3[10]], iq3xxs_grid[q3[11]]); + const uint32x4_t aux32x4_3 = ggml_vld1q_u32(iq3xxs_grid[q3[12]], iq3xxs_grid[q3[13]], iq3xxs_grid[q3[14]], iq3xxs_grid[q3[15]]); + q3 += 16; + q3s.val[0] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[0] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[0] >> 7) & 127)))); + q3s.val[1] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[0] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[0] >> 21) & 127)))); + q3s.val[2] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 0) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 7) & 127)))); + q3s.val[3] = vcombine_s8(vld1_s8((const void *)(signs64 + ((aux32[1] >> 14) & 127))), vld1_s8((const void *)(signs64 + ((aux32[1] >> 21) & 127)))); + q3s.val[0] = vmulq_s8(q3s.val[0], vreinterpretq_s8_u32(aux32x4_0)); + q3s.val[1] = vmulq_s8(q3s.val[1], vreinterpretq_s8_u32(aux32x4_1)); + q3s.val[2] = vmulq_s8(q3s.val[2], vreinterpretq_s8_u32(aux32x4_2)); + q3s.val[3] = vmulq_s8(q3s.val[3], vreinterpretq_s8_u32(aux32x4_3)); + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]); + sumf1 += vaddvq_s32(p1) * (0.5f + (aux32[0] >> 28)); + sumf2 += vaddvq_s32(p2) * (0.5f + (aux32[1] >> 28)); + } + sumf += d*(sumf1 + sumf2); + } + *s = 0.5f * sumf; - const uint8_t * restrict q5 = x[i].qs; +#elif defined(__AVX2__) + + const uint64_t * signs64 = (const uint64_t *)keven_signs_q2xs; + + uint32_t aux32[2]; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q2_1 = _mm256_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], + iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + q3 += 8; + const __m256i q2_2 = _mm256_set_epi32(iq3xxs_grid[q3[7]], iq3xxs_grid[q3[6]], iq3xxs_grid[q3[5]], iq3xxs_grid[q3[4]], + iq3xxs_grid[q3[3]], iq3xxs_grid[q3[2]], iq3xxs_grid[q3[1]], iq3xxs_grid[q3[0]]); + q3 += 8; + memcpy(aux32, gas, 8); gas += 8; + const __m256i s2_1 = _mm256_set_epi64x(signs64[(aux32[0] >> 21) & 127], signs64[(aux32[0] >> 14) & 127], + signs64[(aux32[0] >> 7) & 127], signs64[(aux32[0] >> 0) & 127]); + const __m256i s2_2 = _mm256_set_epi64x(signs64[(aux32[1] >> 21) & 127], signs64[(aux32[1] >> 14) & 127], + signs64[(aux32[1] >> 7) & 127], signs64[(aux32[1] >> 0) & 127]); + const __m256i q8s_1 = _mm256_sign_epi8(q8_1, s2_1); + const __m256i q8s_2 = _mm256_sign_epi8(q8_2, s2_2); + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const uint16_t ls1 = aux32[0] >> 28; + const uint16_t ls2 = aux32[1] >> 28; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = 0.25f * hsum_float_8(accumf); - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); +#else - const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); + uint32_t aux32; - const __m256i scale_l = MM256_SET_M128I(_mm_set1_epi16(x[i].scales[1]), _mm_set1_epi16(x[i].scales[0])); - const __m256i scale_h = MM256_SET_M128I(_mm_set1_epi16(x[i].scales[3]), _mm_set1_epi16(x[i].scales[2])); + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict q3 = x[i].qs; + const uint8_t * restrict gas = x[i].qs + QK_K/4; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ++ib32) { + memcpy(&aux32, gas, sizeof(uint32_t)); gas += sizeof(uint32_t); + const uint32_t ls = 2*(aux32 >> 28) + 1; + int32_t sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3xxs_grid + q3[2*l+0]); + const uint8_t * grid2 = (const uint8_t *)(iq3xxs_grid + q3[2*l+1]); + const uint8_t signs = ksigns_iq2xs[(aux32 >> 7*l) & 127]; + for (int j = 0; j < 4; ++j) { + sumi += grid1[j] * q8[j+0] * (signs & kmask_iq2xs[j+0] ? -1 : 1); + sumi += grid2[j] * q8[j+4] * (signs & kmask_iq2xs[j+4] ? -1 : 1); + } + q8 += 8; + } + q3 += 8; + bsum += sumi * ls; + } + sumf += d * bsum; + } + *s = 0.25f * sumf; +#endif +} - int64_t aux64; - memcpy(&aux64, x[i].qh, 8); - const __m128i haux128 = _mm_set_epi64x(aux64 >> 1, aux64); - const __m256i haux256 = MM256_SET_M128I(_mm_srli_epi16(haux128, 2), haux128); +void ggml_vec_dot_iq3_s_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const __m256i q5h_0 = _mm256_slli_epi16(_mm256_andnot_si256(haux256, mone), 4); - const __m256i q5h_1 = _mm256_slli_epi16(_mm256_andnot_si256(_mm256_srli_epi16(haux256, 4), mone), 4); + const block_iq3_s * restrict x = vx; + const block_q8_K * restrict y = vy; - const __m256i q5l_0 = _mm256_and_si256(q5bits, m4); - const __m256i q5l_1 = _mm256_and_si256(_mm256_srli_epi16(q5bits, 4), m4); + const int nb = n / QK_K; - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); +#if defined(__ARM_NEON) - const __m256i p16_0 = _mm256_madd_epi16(scale_l, _mm256_maddubs_epi16(q5l_0, q8_0)); - const __m256i p16_1 = _mm256_madd_epi16(scale_h, _mm256_maddubs_epi16(q5l_1, q8_1)); - const __m256i s16_0 = _mm256_madd_epi16(scale_l, _mm256_maddubs_epi16(q5h_0, q8_0)); - const __m256i s16_1 = _mm256_madd_epi16(scale_h, _mm256_maddubs_epi16(q5h_1, q8_1)); + typedef union { + uint16x8_t vec_index; + uint16_t index[8]; + } vec_index_t; - const __m256i dot = _mm256_sub_epi32(_mm256_add_epi32(p16_0, p16_1), _mm256_add_epi32(s16_0, s16_1)); + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; - acc = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(dot), acc); + static const uint8_t k_mask2[16] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80,}; - } + static const int16_t k_shift[8] = {8, 7, 6, 5, 4, 3, 2, 1}; - *s = hsum_float_8(acc); + const ggml_uint8x16x2_t mask1 = ggml_vld1q_u8_x2(k_mask1); + const uint8x16_t mask2 = vld1q_u8(k_mask2); -#elif defined __AVX__ + const int16x8_t hshift = vld1q_s16(k_shift); + const uint16x8_t m256 = vdupq_n_u16(256); + const uint8x16_t m1 = vdupq_n_u8(1); - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i mone = _mm_set1_epi8(1); + uint8x16x2_t vs; + ggml_int8x16x4_t q3s; + ggml_int8x16x4_t q8b; + vec_index_t idx; - __m256 acc = _mm256_setzero_ps(); +#if QK_K == 256 + uint32_t scales32[2]; + const uint8_t * scales8 = (const uint8_t *)scales32; +#endif + float sumf = 0; for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)x[i].signs; + const int8_t * restrict q8 = y[i].qs; - const uint8_t * restrict q5 = x[i].qs; - const int8_t * restrict q8 = y[i].qs; +#if QK_K == 256 + memcpy(scales32, x[i].scales, 4); + scales32[1] = (((scales32[0] >> 4) & 0x0f0f0f0f) << 1) | 0x01010101; + scales32[0] = ((scales32[0] & 0x0f0f0f0f) << 1) | 0x01010101; +#endif - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + int sumi1 = 0, sumi2 = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; + + const uint8x16_t idx_l = vld1q_u8(qs); qs += 16; + idx.vec_index = vorrq_u16(vmovl_u8(vget_low_u8 (idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+0]), hshift), m256)); + const uint32x4_t aux32x4_0 = ggml_vld1q_u32(iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]], + iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]); + const uint32x4_t aux32x4_1 = ggml_vld1q_u32(iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]], + iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]); + idx.vec_index = vorrq_u16(vmovl_u8(vget_high_u8(idx_l)), vandq_u16(vshlq_u16(vdupq_n_u16(qh[ib32+1]), hshift), m256)); + const uint32x4_t aux32x4_2 = ggml_vld1q_u32(iq3s_grid[idx.index[0]], iq3s_grid[idx.index[1]], + iq3s_grid[idx.index[2]], iq3s_grid[idx.index[3]]); + const uint32x4_t aux32x4_3 = ggml_vld1q_u32(iq3s_grid[idx.index[4]], iq3s_grid[idx.index[5]], + iq3s_grid[idx.index[6]], iq3s_grid[idx.index[7]]); + + + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[0] | ((uint32_t) signs[1] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1); + vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1); + + q3s.val[0] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_0)); + q3s.val[1] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_1)); + + vs.val[0] = vreinterpretq_u8_u32(vdupq_n_u32(signs[2] | ((uint32_t) signs[3] << 16))); + vs.val[1] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[1]), mask2); + vs.val[0] = vandq_u8(ggml_vqtbl1q_u8(vs.val[0], mask1.val[0]), mask2); + vs.val[0] = vorrq_u8(vceqq_u8(vs.val[0], mask2), m1); + vs.val[1] = vorrq_u8(vceqq_u8(vs.val[1], mask2), m1); + + signs += 4; + + q3s.val[2] = vmulq_s8(vreinterpretq_s8_u8(vs.val[0]), vreinterpretq_s8_u32(aux32x4_2)); + q3s.val[3] = vmulq_s8(vreinterpretq_s8_u8(vs.val[1]), vreinterpretq_s8_u32(aux32x4_3)); + + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[0], q8b.val[0]), q3s.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q3s.val[2], q8b.val[2]), q3s.val[3], q8b.val[3]); +#if QK_K == 256 + sumi1 += vaddvq_s32(p1) * scales8[ib32/2+0]; + sumi2 += vaddvq_s32(p2) * scales8[ib32/2+4]; +#else + sumi1 += vaddvq_s32(p1) * (1 + 2*(x[i].scales[ib32/2] & 0xf)); + sumi2 += vaddvq_s32(p2) * (1 + 2*(x[i].scales[ib32/2] >> 4)); +#endif + } + sumf += d*(sumi1 + sumi2); + } + *s = sumf; - const __m256i q5bits = _mm256_loadu_si256((const __m256i*)q5); +#elif defined(__AVX2__) - const __m128i scale_0 = _mm_set1_epi16(x[i].scales[0]); - const __m128i scale_1 = _mm_set1_epi16(x[i].scales[1]); - const __m128i scale_2 = _mm_set1_epi16(x[i].scales[2]); - const __m128i scale_3 = _mm_set1_epi16(x[i].scales[3]); + static const uint8_t k_mask1[32] = {0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, 0x01, + 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03, 0x03 + }; - int64_t aux64; - memcpy(&aux64, x[i].qh, 8); - const __m128i haux128_0 = _mm_set_epi64x(aux64 >> 1, aux64); - const __m128i haux128_1 = _mm_srli_epi16(haux128_0, 2); + static const uint8_t k_mask2[32] = {0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, 0x01, 0x02, 0x04, 0x08, 0x10, 0x20, 0x40, 0x80, + }; - const __m128i q5h_0 = _mm_slli_epi16(_mm_andnot_si128(haux128_0, mone), 4); - const __m128i q5h_1 = _mm_slli_epi16(_mm_andnot_si128(haux128_1, mone), 4); - const __m128i q5h_2 = _mm_slli_epi16(_mm_andnot_si128(_mm_srli_epi16(haux128_0, 4), mone), 4); - const __m128i q5h_3 = _mm_slli_epi16(_mm_andnot_si128(_mm_srli_epi16(haux128_1, 4), mone), 4); + const __m256i mask1 = _mm256_loadu_si256((const __m256i*)k_mask1); + const __m256i mask2 = _mm256_loadu_si256((const __m256i*)k_mask2); - const __m128i q5l_0 = _mm_and_si128(_mm256_extractf128_si256(q5bits, 0), m4); - const __m128i q5l_1 = _mm_and_si128(_mm256_extractf128_si256(q5bits, 1), m4); - const __m128i q5l_2 = _mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q5bits, 0), 4), m4); - const __m128i q5l_3 = _mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q5bits, 1), 4), m4); + const __m256i idx_shift = _mm256_set_epi32(1, 2, 3, 4, 5, 6, 7, 8); + const __m256i idx_mask = _mm256_set1_epi32(256); - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + typedef union { + __m256i vec[2]; + uint32_t index[16]; + } index_t; - const __m128i p16_0 = _mm_madd_epi16(scale_0, _mm_maddubs_epi16(q5l_0, _mm256_extractf128_si256(q8_0, 0))); - const __m128i p16_1 = _mm_madd_epi16(scale_1, _mm_maddubs_epi16(q5l_1, _mm256_extractf128_si256(q8_0, 1))); - const __m128i p16_2 = _mm_madd_epi16(scale_2, _mm_maddubs_epi16(q5l_2, _mm256_extractf128_si256(q8_1, 0))); - const __m128i p16_3 = _mm_madd_epi16(scale_3, _mm_maddubs_epi16(q5l_3, _mm256_extractf128_si256(q8_1, 1))); - const __m128i s16_0 = _mm_madd_epi16(scale_0, _mm_maddubs_epi16(q5h_0, _mm256_extractf128_si256(q8_0, 0))); - const __m128i s16_1 = _mm_madd_epi16(scale_1, _mm_maddubs_epi16(q5h_1, _mm256_extractf128_si256(q8_0, 1))); - const __m128i s16_2 = _mm_madd_epi16(scale_2, _mm_maddubs_epi16(q5h_2, _mm256_extractf128_si256(q8_1, 0))); - const __m128i s16_3 = _mm_madd_epi16(scale_3, _mm_maddubs_epi16(q5h_3, _mm256_extractf128_si256(q8_1, 1))); + index_t idx; + + __m256 accumf = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint16_t * restrict signs = (const uint16_t *)x[i].signs; + const int8_t * restrict q8 = y[i].qs; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const __m256i q8_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i idx_l = _mm256_cvtepu8_epi16(_mm_loadu_si128((const __m128i *)qs)); qs += 16; + idx.vec[0] = _mm256_set1_epi32(qh[ib32+0]); + idx.vec[1] = _mm256_set1_epi32(qh[ib32+1]); + idx.vec[0] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[0], idx_shift), idx_mask); + idx.vec[1] = _mm256_and_si256(_mm256_sllv_epi32(idx.vec[1], idx_shift), idx_mask); + idx.vec[0] = _mm256_or_si256(idx.vec[0], _mm256_cvtepi16_epi32(_mm256_castsi256_si128(idx_l))); + idx.vec[1] = _mm256_or_si256(idx.vec[1], _mm256_cvtepi16_epi32(_mm256_extractf128_si256(idx_l, 1))); + + // At leat on my CPU (Ryzen 7950X), using _mm256_i32gather_epi32 is slower than _mm256_set_epi32. Strange. + //const __m256i q2_1 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[0], 4); + //const __m256i q2_2 = _mm256_i32gather_epi32((const int *)iq3s_grid, idx.vec[1], 4); + const __m256i q2_1 = _mm256_set_epi32( + iq3s_grid[idx.index[7]], iq3s_grid[idx.index[6]], iq3s_grid[idx.index[5]], iq3s_grid[idx.index[4]], + iq3s_grid[idx.index[3]], iq3s_grid[idx.index[2]], iq3s_grid[idx.index[1]], iq3s_grid[idx.index[0]] + ); + const __m256i q2_2 = _mm256_set_epi32( + iq3s_grid[idx.index[15]], iq3s_grid[idx.index[14]], iq3s_grid[idx.index[13]], iq3s_grid[idx.index[12]], + iq3s_grid[idx.index[11]], iq3s_grid[idx.index[10]], iq3s_grid[idx.index[ 9]], iq3s_grid[idx.index[ 8]] + ); + + __m256i aux256 = _mm256_set1_epi32(signs[0] | (signs[1] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_1 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_1 = _mm256_sub_epi8(_mm256_xor_si256(s2_1, q8_1), s2_1); + + aux256 = _mm256_set1_epi32(signs[2] | (signs[3] << 16)); + aux256 = _mm256_and_si256(_mm256_shuffle_epi8(aux256,mask1), mask2); + const __m256i s2_2 = _mm256_cmpeq_epi8(aux256, mask2); + const __m256i q8s_2 = _mm256_sub_epi8(_mm256_xor_si256(s2_2, q8_2), s2_2); + + signs += 4; + + const __m256i dot1 = _mm256_maddubs_epi16(q2_1, q8s_1); + const __m256i dot2 = _mm256_maddubs_epi16(q2_2, q8s_2); + const uint16_t ls1 = x[i].scales[ib32/2] & 0xf; + const uint16_t ls2 = x[i].scales[ib32/2] >> 4; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(2*ls1+1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(2*ls2+1)); + sumi1 = _mm256_add_epi32(sumi1, p1); + sumi2 = _mm256_add_epi32(sumi2, p2); + } + + accumf = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accumf); + + } + + *s = hsum_float_8(accumf); + +#else + + float sumf = 0.f; + for (int i = 0; i < nb; ++i) { + const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + const uint8_t * restrict qs = x[i].qs; + const uint8_t * restrict qh = x[i].qh; + const uint8_t * restrict signs = x[i].signs; + const int8_t * restrict q8 = y[i].qs; + int32_t bsum = 0; + for (int ib32 = 0; ib32 < QK_K/32; ib32 += 2) { + const uint32_t ls1 = 2*(x[i].scales[ib32/2] & 0xf) + 1; + const uint32_t ls2 = 2*(x[i].scales[ib32/2] >> 4) + 1; + int32_t sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+0] << (8-2*l)) & 256))); + const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+0] << (7-2*l)) & 256))); + for (int j = 0; j < 4; ++j) { + sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1); + sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1); + } + q8 += 8; + } + qs += 8; + signs += 4; + bsum += sumi * ls1; + sumi = 0; + for (int l = 0; l < 4; ++l) { + const uint8_t * grid1 = (const uint8_t *)(iq3s_grid + (qs[2*l+0] | ((qh[ib32+1] << (8-2*l)) & 256))); + const uint8_t * grid2 = (const uint8_t *)(iq3s_grid + (qs[2*l+1] | ((qh[ib32+1] << (7-2*l)) & 256))); + for (int j = 0; j < 4; ++j) { + sumi += grid1[j] * q8[j+0] * (signs[l] & kmask_iq2xs[j+0] ? -1 : 1); + sumi += grid2[j] * q8[j+4] * (signs[l] & kmask_iq2xs[j+4] ? -1 : 1); + } + q8 += 8; + } + qs += 8; + signs += 4; + bsum += sumi * ls2; + } + sumf += d * bsum; + } + *s = sumf; +#endif +} - const __m128i dot_0 = _mm_sub_epi32(_mm_add_epi32(p16_0, p16_2), _mm_add_epi32(s16_0, s16_2)); - const __m128i dot_1 = _mm_sub_epi32(_mm_add_epi32(p16_1, p16_3), _mm_add_epi32(s16_1, s16_3)); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(MM256_SET_M128I(dot_1, dot_0))), acc); +#ifdef __AVX2__ +static inline __m256i mul_add_epi8(const __m256i x, const __m256i y) { + const __m256i ax = _mm256_sign_epi8(x, x); + const __m256i sy = _mm256_sign_epi8(y, x); + return _mm256_maddubs_epi16(ax, sy); +} +#endif - } +void ggml_vec_dot_iq1_s_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - *s = hsum_float_8(acc); + const block_iq1_s * restrict x = vx; + const block_q8_K * restrict y = vy; -#elif defined __riscv_v_intrinsic + const int nb = n / QK_K; - float sumf = 0; +#if defined __ARM_NEON + + ggml_int8x16x4_t q1b; + ggml_int8x16x4_t q8b; + float sumf = 0; for (int i = 0; i < nb; ++i) { - const float d = y[i].d * (float)x[i].d; - const int8_t * sc = x[i].scales; + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; - const uint8_t * restrict q5 = x[i].qs; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; + int sumi1 = 0, sumi2 = 0, sumi3 = 0; - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + for (int ib = 0; ib < QK_K/32; ib += 2) { - // load qh - vuint8mf4_t qh_x1 = __riscv_vle8_v_u8mf4(qh, 8); - vuint8mf2_t qh_x2 = __riscv_vlmul_ext_v_u8mf4_u8mf2(__riscv_vsrl_vx_u8mf4(qh_x1, 1, 8)); + q1b.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[0] | ((qh[ib+0] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[1] | ((qh[ib+0] << 5) & 0x700))))); + q1b.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[2] | ((qh[ib+0] << 2) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[3] | ((qh[ib+0] >> 1) & 0x700))))); + q1b.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[4] | ((qh[ib+1] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[5] | ((qh[ib+1] << 5) & 0x700))))); + q1b.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[6] | ((qh[ib+1] << 2) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[7] | ((qh[ib+1] >> 1) & 0x700))))); + qs += 8; - size_t vl = 16; + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - // combine both qh_1 and qh_2 - vuint8mf2_t qh_x = __riscv_vslideup_vx_u8mf2(__riscv_vlmul_ext_v_u8mf4_u8mf2(qh_x1), qh_x2, vl/2, vl); + const int32x4_t p1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q1b.val[0], q8b.val[0]), q1b.val[1], q8b.val[1]); + const int32x4_t p2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q1b.val[2], q8b.val[2]), q1b.val[3], q8b.val[3]); - vuint8mf2_t qh_h0 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(__riscv_vsll_vx_u8mf2(qh_x, 0x4, vl), vl), 16, vl); - vuint8mf2_t qh_h1 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(__riscv_vsll_vx_u8mf2(qh_x, 0x2, vl), vl), 16, vl); - vuint8mf2_t qh_h2 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(qh_x, vl), 16, vl); - vuint8mf2_t qh_h3 = __riscv_vand_vx_u8mf2(__riscv_vnot_v_u8mf2(__riscv_vsrl_vx_u8mf2(qh_x, 0x4, vl), vl), 16, vl); + const int ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; + const int ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; + sumi1 += vaddvq_s32(p1) * ls1; + sumi2 += vaddvq_s32(p2) * ls2; + sumi3 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * ls1 * (qh[ib+0] & 0x8000 ? -1 : 1) + + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * ls2 * (qh[ib+1] & 0x8000 ? -1 : 1); - vint8mf2_t qh_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h0); - vint8mf2_t qh_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h1); - vint8mf2_t qh_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h2); - vint8mf2_t qh_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(qh_h3); + } - // load q5 - vuint8mf2_t q5_x1 = __riscv_vle8_v_u8mf2(q5, vl); - vuint8mf2_t q5_x2 = __riscv_vle8_v_u8mf2(q5+16, vl); + sumf += y[i].d * GGML_FP16_TO_FP32(x[i].d) * (sumi1 + sumi2 + IQ1S_DELTA * sumi3); + } - vint8mf2_t q5s_0 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(q5_x1, 0xF, vl)); - vint8mf2_t q5s_1 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vand_vx_u8mf2(q5_x2, 0xF, vl)); - vint8mf2_t q5s_2 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vsrl_vx_u8mf2(q5_x1, 0x4, vl)); - vint8mf2_t q5s_3 = __riscv_vreinterpret_v_u8mf2_i8mf2(__riscv_vsrl_vx_u8mf2(q5_x2, 0x4, vl)); + *s = sumf; - vint8mf2_t q5_0 = __riscv_vsub_vv_i8mf2(q5s_0, qh_0, vl); - vint8mf2_t q5_1 = __riscv_vsub_vv_i8mf2(q5s_1, qh_1, vl); - vint8mf2_t q5_2 = __riscv_vsub_vv_i8mf2(q5s_2, qh_2, vl); - vint8mf2_t q5_3 = __riscv_vsub_vv_i8mf2(q5s_3, qh_3, vl); +#elif defined __AVX2__ - // load Q8 and multiply it with Q5 - vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q5_0, __riscv_vle8_v_i8mf2(q8, vl), vl); - vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q5_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); - vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q5_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); - vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q5_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); + __m256 accum = _mm256_setzero_ps(); + float accum1 = 0; + for (int i = 0; i < nb; ++i) { - vint32m1_t vs_0 = __riscv_vwredsum_vs_i16m1_i32m1(p0, vzero, vl); - vint32m1_t vs_1 = __riscv_vwredsum_vs_i16m1_i32m1(p1, vzero, vl); - vint32m1_t vs_2 = __riscv_vwredsum_vs_i16m1_i32m1(p2, vzero, vl); - vint32m1_t vs_3 = __riscv_vwredsum_vs_i16m1_i32m1(p3, vzero, vl); + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; - int32_t sumi1 = sc[0] * __riscv_vmv_x_s_i32m1_i32(vs_0); - int32_t sumi2 = sc[1] * __riscv_vmv_x_s_i32m1_i32(vs_1); - int32_t sumi3 = sc[2] * __riscv_vmv_x_s_i32m1_i32(vs_2); - int32_t sumi4 = sc[3] * __riscv_vmv_x_s_i32m1_i32(vs_3); + __m256i sumi = _mm256_setzero_si256(); + int sumi1 = 0; + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m256i q1b_1 = _mm256_set_epi64x(iq1s_grid[qs[3] | ((qh[ib+0] >> 1) & 0x700)], iq1s_grid[qs[2] | ((qh[ib+0] << 2) & 0x700)], + iq1s_grid[qs[1] | ((qh[ib+0] << 5) & 0x700)], iq1s_grid[qs[0] | ((qh[ib+0] << 8) & 0x700)]); + const __m256i q1b_2 = _mm256_set_epi64x(iq1s_grid[qs[7] | ((qh[ib+1] >> 1) & 0x700)], iq1s_grid[qs[6] | ((qh[ib+1] << 2) & 0x700)], + iq1s_grid[qs[5] | ((qh[ib+1] << 5) & 0x700)], iq1s_grid[qs[4] | ((qh[ib+1] << 8) & 0x700)]); + qs += 8; + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); + const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); + const int16_t ls1 = 2*((qh[ib+0] >> 12) & 7) + 1; + const int16_t ls2 = 2*((qh[ib+1] >> 12) & 7) + 1; + const __m256i p1 = _mm256_madd_epi16(dot1, _mm256_set1_epi16(ls1)); + const __m256i p2 = _mm256_madd_epi16(dot2, _mm256_set1_epi16(ls2)); + + sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p1, p2)); + sumi1 += (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]) * (qh[ib+0] & 0x8000 ? -1 : 1) * ls1 + + (y[i].bsums[2*ib+2] + y[i].bsums[2*ib+3]) * (qh[ib+1] & 0x8000 ? -1 : 1) * ls2; + } - sumf += d * (sumi1 + sumi2 + sumi3 + sumi4); + const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + accum = _mm256_fmadd_ps(_mm256_set1_ps(d), _mm256_cvtepi32_ps(sumi), accum); + accum1 += d * sumi1; } - *s = sumf; + *s = hsum_float_8(accum) + IQ1S_DELTA * accum1; #else - int8_t aux8[QK_K]; - int16_t aux16[16]; - float sums [8]; - memset(sums, 0, 8*sizeof(float)); - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].qs; - const uint8_t * restrict hm = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - int8_t * restrict a = aux8; - for (int l = 0; l < 32; ++l) { - a[l+ 0] = q4[l] & 0xF; - a[l+32] = q4[l] >> 4; - } - for (int is = 0; is < 8; ++is) { - uint8_t m = 1 << is; - for (int l = 0; l < 8; ++l) a[8*is + l] -= (hm[l] & m ? 0 : 16); - } - - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); - const int8_t * restrict sc = x[i].scales; + for (int i = 0; i < nb; i++) { - for (int j = 0; j < QK_K/16; ++j) { - const float dl = d * sc[j]; - for (int l = 0; l < 16; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) sums[l] += dl * (aux16[l] + aux16[8+l]); - q8 += 16; a += 16; + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint16_t * qh = x[i].qh; + + int sumi = 0, sumi1 = 0; + for (int ib = 0; ib < QK_K/32; ++ib) { + const int ls = 2*((qh[ib] >> 12) & 7) + 1; + const int delta = qh[ib] & 0x8000 ? -1 : 1; + int lsum = 0; + for (int l = 0; l < 4; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((qh[ib] >> 3*l) & 7) << 8))); + for (int j = 0; j < 8; ++j) { + lsum += q8[j] * grid[j]; + } + q8 += 8; + } + sumi += ls * lsum; + sumi1 += ls * delta * (y[i].bsums[2*ib+0] + y[i].bsums[2*ib+1]); + qs += 4; } + + sumf += GGML_FP16_TO_FP32(x[i].d) * y[i].d * (sumi + IQ1S_DELTA * sumi1); } - for (int l = 0; l < 8; ++l) sumf += sums[l]; + *s = sumf; + #endif } -#endif - -#if QK_K == 256 -void ggml_vec_dot_q6_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { +void ggml_vec_dot_iq1_m_q8_K (int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { assert(n % QK_K == 0); + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); - const block_q6_K * restrict x = vx; - const block_q8_K * restrict y = vy; + const block_iq1_m * restrict x = vx; + const block_q8_K * restrict y = vy; const int nb = n / QK_K; -#ifdef __ARM_NEON +#if QK_K != 64 + iq1m_scale_t scale; +#endif - float sum = 0; +#if defined __ARM_NEON - const uint8x16_t m4b = vdupq_n_u8(0xF); -#if defined(__ARM_FEATURE_DOTPROD) - const int32x4_t vzero = vdupq_n_s32(0); +#if QK_K == 64 + const int32x4_t mask = vdupq_n_s32(0xf); +#else + const int32x4_t mask = vdupq_n_s32(0x7); #endif - //const int8x16_t m32s = vdupq_n_s8(32); + const int32x4_t mone = vdupq_n_s32(1); + const int32x4_t mzero = vdupq_n_s32(0); - const uint8x16_t mone = vdupq_n_u8(3); + ggml_int8x16x4_t deltas; + deltas.val[0] = vcombine_s8(vdup_n_s8(+1), vdup_n_s8(+1)); + deltas.val[1] = vcombine_s8(vdup_n_s8(-1), vdup_n_s8(+1)); + deltas.val[2] = vcombine_s8(vdup_n_s8(+1), vdup_n_s8(-1)); + deltas.val[3] = vcombine_s8(vdup_n_s8(-1), vdup_n_s8(-1)); + + ggml_int8x16x4_t q1b; + ggml_int8x16x4_t q8b; - int8x16x4_t q6bytes; - uint8x16x4_t q6h; + uint32_t aux32; + const uint8_t * aux8 = (const uint8_t *)&aux32; + float sumf = 0; for (int i = 0; i < nb; ++i) { - const float d_all = GGML_FP16_TO_FP32(x[i].d); + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; +#if QK_K != 64 + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); +#endif - const int8_t * restrict scale = x[i].scales; + int32x4_t sumi1 = mzero; + int32x4_t sumi2 = mzero; - const int16x8x2_t q8sums = vld1q_s16_x2(y[i].bsums); - const int8x16_t scales = vld1q_s8(scale); - const int16x8x2_t q6scales = {vmovl_s8(vget_low_s8(scales)), vmovl_s8(vget_high_s8(scales))}; + for (int ib = 0; ib < QK_K/32; ib += 2) { - const int32x4_t prod = vaddq_s32(vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[0]), vget_low_s16 (q6scales.val[0])), - vmull_s16(vget_high_s16(q8sums.val[0]), vget_high_s16(q6scales.val[0]))), - vaddq_s32(vmull_s16(vget_low_s16 (q8sums.val[1]), vget_low_s16 (q6scales.val[1])), - vmull_s16(vget_high_s16(q8sums.val[1]), vget_high_s16(q6scales.val[1])))); - int32_t isum_mins = vaddvq_s32(prod); + q1b.val[0] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[0] | ((qh[0] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[1] | ((qh[0] << 4) & 0x700))))); + q1b.val[1] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[2] | ((qh[1] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[3] | ((qh[1] << 4) & 0x700))))); + q1b.val[2] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[4] | ((qh[2] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[5] | ((qh[2] << 4) & 0x700))))); + q1b.val[3] = vcombine_s8(vld1_s8((const int8_t *)(iq1s_grid + (qs[6] | ((qh[3] << 8) & 0x700)))), + vld1_s8((const int8_t *)(iq1s_grid + (qs[7] | ((qh[3] << 4) & 0x700))))); - int32_t isum = 0; + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - for (int j = 0; j < QK_K/128; ++j) { + const int32x4_t p1 = vpaddq_s32(ggml_vdotq_s32(mzero, q1b.val[0], q8b.val[0]), ggml_vdotq_s32(mzero, q1b.val[1], q8b.val[1])); + const int32x4_t p2 = vpaddq_s32(ggml_vdotq_s32(mzero, q1b.val[2], q8b.val[2]), ggml_vdotq_s32(mzero, q1b.val[3], q8b.val[3])); + const int32x4_t p12 = vpaddq_s32(p1, p2); - uint8x16x2_t qhbits = vld1q_u8_x2(qh); qh += 32; - uint8x16x4_t q6bits = vld1q_u8_x4(q6); q6 += 64; - int8x16x4_t q8bytes = vld1q_s8_x4(q8); q8 += 64; + const uint32_t * qh32 = (const uint32_t *)qh; // we are 4-byte aligned, so we can do that + aux32 = ((qh32[0] >> 3) & 0x01010101) | ((qh32[0] >> 6) & 0x02020202); - q6h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits.val[0]), 4); - q6h.val[1] = vshlq_n_u8(vandq_u8(mone, qhbits.val[1]), 4); - uint8x16_t shifted = vshrq_n_u8(qhbits.val[0], 2); - q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[1], 2); - q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + const int32x4_t p3 = vpaddq_s32(ggml_vdotq_s32(mzero, deltas.val[aux8[0]], q8b.val[0]), ggml_vdotq_s32(mzero, deltas.val[aux8[1]], q8b.val[1])); + const int32x4_t p4 = vpaddq_s32(ggml_vdotq_s32(mzero, deltas.val[aux8[2]], q8b.val[2]), ggml_vdotq_s32(mzero, deltas.val[aux8[3]], q8b.val[3])); + const int32x4_t p34 = vpaddq_s32(p3, p4); - //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])), m32s); - //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])), m32s); - //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])), m32s); - //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])), m32s); - q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])); - q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])); - q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[2], m4b), q6h.val[2])); - q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[3], m4b), q6h.val[3])); +#if QK_K == 64 + int32x4_t scales_4 = ggml_vld1q_u32(sc[0] >> 0, sc[0] >> 4, sc[0] >> 8, sc[0] >> 12); +#else + int32x4_t scales_4 = ggml_vld1q_u32(sc[ib/2] >> 0, sc[ib/2] >> 3, sc[ib/2] >> 6, sc[ib/2] >> 9); +#endif + scales_4 = vaddq_s32(vshlq_n_s32(vandq_s32(scales_4, mask), 1), mone); -#if defined(__ARM_FEATURE_DOTPROD) + sumi1 = vmlaq_s32(sumi1, scales_4, p12); + sumi2 = vmlaq_s32(sumi2, scales_4, p34); - isum += vaddvq_s32(vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; - scale += 4; + qs += 8; qh += 4; + + } + +#if QK_K == 64 + sumf += y[i].d * GGML_FP16_TO_FP32(x[i].d) * (vaddvq_s32(sumi1) + IQ1M_DELTA * vaddvq_s32(sumi2)); +#else + sumf += y[i].d * GGML_FP16_TO_FP32(scale.f16) * (vaddvq_s32(sumi1) + IQ1M_DELTA * vaddvq_s32(sumi2)); +#endif + } + + *s = sumf; +#elif defined __AVX2__ + +#if QK_K == 64 + const __m256i mask = _mm256_set1_epi16(0xf); #else + const __m256i mask = _mm256_set1_epi16(0x7); +#endif + const __m256i mone = _mm256_set1_epi16(1); + + __m256 accum1 = _mm256_setzero_ps(); + __m256 accum2 = _mm256_setzero_ps(); + for (int i = 0; i < nb; ++i) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; - int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q6bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q6bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - isum += vaddvq_s16(p0) * scale[0] + vaddvq_s16(p1) * scale[1]; - scale += 2; - - int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q6bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q6bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - isum += vaddvq_s16(p2) * scale[0] + vaddvq_s16(p3) * scale[1]; - scale += 2; +#if QK_K != 64 + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); #endif - q8bytes = vld1q_s8_x4(q8); q8 += 64; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m256i q1b_1 = _mm256_set_epi64x( + iq1s_grid[qs[3] | (((uint16_t)qh[1] << 4) & 0x700)], iq1s_grid[qs[2] | (((uint16_t)qh[1] << 8) & 0x700)], + iq1s_grid[qs[1] | (((uint16_t)qh[0] << 4) & 0x700)], iq1s_grid[qs[0] | (((uint16_t)qh[0] << 8) & 0x700)] + ); + const __m256i q1b_2 = _mm256_set_epi64x( + iq1s_grid[qs[7] | (((uint16_t)qh[3] << 4) & 0x700)], iq1s_grid[qs[6] | (((uint16_t)qh[3] << 8) & 0x700)], + iq1s_grid[qs[5] | (((uint16_t)qh[2] << 4) & 0x700)], iq1s_grid[qs[4] | (((uint16_t)qh[2] << 8) & 0x700)] + ); + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + + const __m256i dot1 = mul_add_epi8(q1b_1, q8b_1); + const __m256i dot2 = mul_add_epi8(q1b_2, q8b_2); + + const __m256i delta1 = _mm256_set_epi64x(qh[1] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[1] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101, + qh[0] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[0] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + const __m256i delta2 = _mm256_set_epi64x(qh[3] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[3] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101, + qh[2] & 0x80 ? 0xffffffffffffffff : 0x0101010101010101, + qh[2] & 0x08 ? 0xffffffffffffffff : 0x0101010101010101); + + const __m256i dot3 = mul_add_epi8(delta1, q8b_1); + const __m256i dot4 = mul_add_epi8(delta2, q8b_2); +#if QK_K == 64 + __m256i scale1 = MM256_SET_M128I(_mm_set1_epi16(sc[0] >> 4), _mm_set1_epi16(sc[0] >> 0)); + __m256i scale2 = MM256_SET_M128I(_mm_set1_epi16(sc[0] >> 12), _mm_set1_epi16(sc[0] >> 8)); +#else + __m256i scale1 = MM256_SET_M128I(_mm_set1_epi16(sc[ib/2] >> 3), _mm_set1_epi16(sc[ib/2] >> 0)); + __m256i scale2 = MM256_SET_M128I(_mm_set1_epi16(sc[ib/2] >> 9), _mm_set1_epi16(sc[ib/2] >> 6)); +#endif + scale1 = _mm256_add_epi16(_mm256_slli_epi16(_mm256_and_si256(scale1, mask), 1), mone); + scale2 = _mm256_add_epi16(_mm256_slli_epi16(_mm256_and_si256(scale2, mask), 1), mone); + const __m256i p1 = _mm256_madd_epi16(dot1, scale1); + const __m256i p2 = _mm256_madd_epi16(dot2, scale2); + const __m256i p3 = _mm256_madd_epi16(dot3, scale1); + const __m256i p4 = _mm256_madd_epi16(dot4, scale2); - shifted = vshrq_n_u8(qhbits.val[0], 4); - q6h.val[0] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[1], 4); - q6h.val[1] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[0], 6); - q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits.val[1], 6); - q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + sumi1 = _mm256_add_epi32(sumi1, _mm256_add_epi32(p1, p2)); + sumi2 = _mm256_add_epi32(sumi2, _mm256_add_epi32(p3, p4)); - //q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])), m32s); - //q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])), m32s); - //q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])), m32s); - //q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])), m32s); - q6bytes.val[0] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[0])); - q6bytes.val[1] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[1])); - q6bytes.val[2] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[2], 4), q6h.val[2])); - q6bytes.val[3] = vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[3], 4), q6h.val[3])); + qs += 8; qh += 4; + } -#if defined(__ARM_FEATURE_DOTPROD) +#if QK_K == 64 + const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(x[i].d)); +#else + const __m256 d = _mm256_set1_ps(y[i].d * GGML_FP16_TO_FP32(scale.f16)); +#endif + accum1 = _mm256_fmadd_ps(d, _mm256_cvtepi32_ps(sumi1), accum1); + accum2 = _mm256_fmadd_ps(d, _mm256_cvtepi32_ps(sumi2), accum2); - isum += vaddvq_s32(vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; - scale += 4; + } + + *s = hsum_float_8(accum1) + IQ1M_DELTA * hsum_float_8(accum2); - //for (int l = 0; l < 4; ++l) { - // const int32x4_t p = vdotq_s32(vzero, q6bytes.val[l], q8bytes.val[l]); - // isum += vaddvq_s32(p) * *scale++; - //} #else - p0 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q6bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - p1 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q6bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - isum += vaddvq_s16(p0) * scale[0] + vaddvq_s16(p1) * scale[1]; - scale += 2; - - p2 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q6bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - p3 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q6bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - isum += vaddvq_s16(p2) * scale[0] + vaddvq_s16(p3) * scale[1]; - scale += 2; + + int sum1[2], sum2[2], delta[4]; + + float sumf = 0; + for (int i = 0; i < nb; i++) { + + const int8_t * q8 = y[i].qs; + const uint8_t * qs = x[i].qs; + const uint8_t * qh = x[i].qh; + const uint16_t * sc = (const uint16_t *)x[i].scales; + +#if QK_K != 64 + scale.u16 = (sc[0] >> 12) | ((sc[1] >> 8) & 0x00f0) | ((sc[2] >> 4) & 0x0f00) | (sc[3] & 0xf000); #endif + int sumi1 = 0, sumi2 = 0; + for (int ib = 0; ib < QK_K/32; ++ib) { + delta[0] = qh[0] & 0x08 ? -1 : 1; + delta[1] = qh[0] & 0x80 ? -1 : 1; + delta[2] = qh[1] & 0x08 ? -1 : 1; + delta[3] = qh[1] & 0x80 ? -1 : 1; + sum1[0] = sum1[1] = sum2[0] = sum2[1] = 0; + for (int l = 0; l < 4; ++l) { + const int8_t * grid = (const int8_t *)(iq1s_grid + (qs[l] | (((uint16_t)qh[l/2] << (8 - 4*(l%2))) & 0x700))); + int lsum1 = 0, lsum2 = 0; + for (int j = 0; j < 8; ++j) { + lsum1 += q8[j] * grid[j]; + lsum2 += q8[j]; + } + q8 += 8; + sum1[l/2] += lsum1; + sum2[l/2] += lsum2*delta[l]; + } +#if QK_K == 64 + const int ls1 = 2*((sc[0] >> (8*(ib%2)+0)) & 0xf) + 1; + const int ls2 = 2*((sc[0] >> (8*(ib%2)+4)) & 0xf) + 1; +#else + const int ls1 = 2*((sc[ib/2] >> (6*(ib%2)+0)) & 0x7) + 1; + const int ls2 = 2*((sc[ib/2] >> (6*(ib%2)+3)) & 0x7) + 1; +#endif + sumi1 += sum1[0] * ls1 + sum1[1] * ls2; + sumi2 += sum2[0] * ls1 + sum2[1] * ls2; + qs += 4; + qh += 2; } - //sum += isum * d_all * y[i].d; - sum += d_all * y[i].d * (isum - 32 * isum_mins); +#if QK_K == 64 + sumf += GGML_FP16_TO_FP32(x[i].d) * y[i].d * (sumi1 + IQ1M_DELTA * sumi2); +#else + sumf += GGML_FP16_TO_FP32(scale.f16) * y[i].d * (sumi1 + IQ1M_DELTA * sumi2); +#endif } - *s = sum; -#elif defined __AVX2__ + *s = sumf; - const __m256i m4 = _mm256_set1_epi8(0xF); - const __m256i m2 = _mm256_set1_epi8(3); - const __m256i m32s = _mm256_set1_epi8(32); +#endif +} - __m256 acc = _mm256_setzero_ps(); +void ggml_vec_dot_iq4_nl_q8_0(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + assert(n % QK4_NL == 0); + static_assert(QK4_NL == QK8_0, "QK4_NL and QK8_0 must be the same"); - for (int i = 0; i < nb; ++i) { + const block_iq4_nl * restrict x = vx; + const block_q8_0 * restrict y = vy; - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + const int nb = n / QK4_NL; - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; +#if defined __ARM_NEON + const int8x16_t values = vld1q_s8(kvalues_iq4nl); + const uint8x16_t m4b = vdupq_n_u8(0x0f); + uint8x16x2_t q4bits; + int8x16x4_t q4b; + int8x16x4_t q8b; + int32x4_t prod_1, prod_2; - const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); + float sumf = 0; - __m256i sumi = _mm256_setzero_si256(); + for (int ib = 0; ib < nb; ib += 2) { - int is = 0; + q4bits.val[0] = vld1q_u8(x[ib+0].qs); + q4bits.val[1] = vld1q_u8(x[ib+1].qs); + q8b.val[0] = vld1q_s8(y[ib+0].qs); + q8b.val[1] = vld1q_s8(y[ib+0].qs + 16); + q8b.val[2] = vld1q_s8(y[ib+1].qs); + q8b.val[3] = vld1q_s8(y[ib+1].qs + 16); - for (int j = 0; j < QK_K/128; ++j) { + q4b.val[0] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[0], m4b)); + q4b.val[1] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[0], 4)); + q4b.val[2] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[1], m4b)); + q4b.val[3] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[1], 4)); - const __m128i scale_0 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 0)); - const __m128i scale_1 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 1)); - const __m128i scale_2 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 2)); - const __m128i scale_3 = _mm_shuffle_epi8(scales, get_scale_shuffle(is + 3)); - is += 4; + prod_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[0], q8b.val[0]), q4b.val[1], q8b.val[1]); + prod_2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[2], q8b.val[2]), q4b.val[3], q8b.val[3]); - const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; - const __m256i q4bits2 = _mm256_loadu_si256((const __m256i*)q4); q4 += 32; - const __m256i q4bitsH = _mm256_loadu_si256((const __m256i*)qh); qh += 32; + sumf += + GGML_FP16_TO_FP32(x[ib+0].d) * GGML_FP16_TO_FP32(y[ib+0].d) * vaddvq_s32(prod_1) + + GGML_FP16_TO_FP32(x[ib+1].d) * GGML_FP16_TO_FP32(y[ib+1].d) * vaddvq_s32(prod_2); + } - const __m256i q4h_0 = _mm256_slli_epi16(_mm256_and_si256(q4bitsH, m2), 4); - const __m256i q4h_1 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 2), m2), 4); - const __m256i q4h_2 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 4), m2), 4); - const __m256i q4h_3 = _mm256_slli_epi16(_mm256_and_si256(_mm256_srli_epi16(q4bitsH, 6), m2), 4); + *s = sumf; + +#elif defined __AVX2__ + + const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); + const __m128i m4b = _mm_set1_epi8(0x0f); + const __m256i mone = _mm256_set1_epi16(1); + + __m256 accum1 = _mm256_setzero_ps(); + __m256 accum2 = _mm256_setzero_ps(); + for (int ib = 0; ib < nb; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i*)x[0].qs); + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)x[1].qs); + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)y[0].qs); + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)y[1].qs); + const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b))); + const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b))); + const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); + const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); + const __m256i p_1 = _mm256_madd_epi16(p16_1, mone); + const __m256i p_2 = _mm256_madd_epi16(p16_2, mone); + accum1 = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[0].d)*GGML_FP16_TO_FP32(x[0].d)), + _mm256_cvtepi32_ps(p_1), accum1); + accum2 = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(y[1].d)*GGML_FP16_TO_FP32(x[1].d)), + _mm256_cvtepi32_ps(p_2), accum2); + + y += 2; + x += 2; + } + + *s = hsum_float_8(_mm256_add_ps(accum1, accum2)); + +#else + float sumf = 0; + for (int ib = 0; ib < nb; ++ib) { + const float d = GGML_FP16_TO_FP32(y[ib].d)*GGML_FP16_TO_FP32(x[ib].d); + int sumi1 = 0, sumi2 = 0; + for (int j = 0; j < QK4_NL/2; ++j) { + sumi1 += y[ib].qs[j+ 0] * kvalues_iq4nl[x[ib].qs[j] & 0xf]; + sumi2 += y[ib].qs[j+QK4_NL/2] * kvalues_iq4nl[x[ib].qs[j] >> 4]; + } + sumf += d * (sumi1 + sumi2); + } + *s = sumf; +#endif +} + +void ggml_vec_dot_iq4_xs_q8_K(int n, float * restrict s, size_t bs, const void * restrict vx, size_t bx, const void * restrict vy, size_t by, int nrc) { + assert(nrc == 1); + UNUSED(nrc); + UNUSED(bx); + UNUSED(by); + UNUSED(bs); + assert(n % QK_K == 0); +#if QK_K == 64 + ggml_vec_dot_iq4_nl_q8_0(n, s, bs, vx, bx, vy, by, nrc); +#else + + const block_iq4_xs * restrict x = vx; + const block_q8_K * restrict y = vy; + + const int nb = n / QK_K; + +#if defined __ARM_NEON + const int8x16_t values = vld1q_s8(kvalues_iq4nl); + const uint8x16_t m4b = vdupq_n_u8(0x0f); + ggml_uint8x16x2_t q4bits; + ggml_int8x16x4_t q4b; + ggml_int8x16x4_t q8b; + int32x4_t prod_1, prod_2; - const __m256i q4_0 = _mm256_or_si256(_mm256_and_si256(q4bits1, m4), q4h_0); - const __m256i q4_1 = _mm256_or_si256(_mm256_and_si256(q4bits2, m4), q4h_1); - const __m256i q4_2 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits1, 4), m4), q4h_2); - const __m256i q4_3 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits2, 4), m4), q4h_3); + float sumf = 0; - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_2 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; - const __m256i q8_3 = _mm256_loadu_si256((const __m256i*)q8); q8 += 32; + for (int ibl = 0; ibl < nb; ++ibl) { - __m256i q8s_0 = _mm256_maddubs_epi16(m32s, q8_0); - __m256i q8s_1 = _mm256_maddubs_epi16(m32s, q8_1); - __m256i q8s_2 = _mm256_maddubs_epi16(m32s, q8_2); - __m256i q8s_3 = _mm256_maddubs_epi16(m32s, q8_3); + const int8_t * q8 = y[ibl].qs; + const uint8_t * q4 = x[ibl].qs; + uint16_t h = x[ibl].scales_h; - __m256i p16_0 = _mm256_maddubs_epi16(q4_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q4_1, q8_1); - __m256i p16_2 = _mm256_maddubs_epi16(q4_2, q8_2); - __m256i p16_3 = _mm256_maddubs_epi16(q4_3, q8_3); + int sumi1 = 0, sumi2 = 0; + for (int ib = 0; ib < QK_K/64; ++ib) { - p16_0 = _mm256_sub_epi16(p16_0, q8s_0); - p16_1 = _mm256_sub_epi16(p16_1, q8s_1); - p16_2 = _mm256_sub_epi16(p16_2, q8s_2); - p16_3 = _mm256_sub_epi16(p16_3, q8s_3); + q4bits = ggml_vld1q_u8_x2(q4); q4 += 32; + q8b = ggml_vld1q_s8_x4(q8); q8 += 64; - p16_0 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_0), p16_0); - p16_1 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_1), p16_1); - p16_2 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_2), p16_2); - p16_3 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_3), p16_3); + q4b.val[0] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[0], m4b)); + q4b.val[1] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[0], 4)); + q4b.val[2] = ggml_vqtbl1q_s8(values, vandq_u8 (q4bits.val[1], m4b)); + q4b.val[3] = ggml_vqtbl1q_s8(values, vshrq_n_u8(q4bits.val[1], 4)); - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_2, p16_3)); + prod_1 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[0], q8b.val[0]), q4b.val[1], q8b.val[1]); + prod_2 = ggml_vdotq_s32(ggml_vdotq_s32(vdupq_n_s32(0), q4b.val[2], q8b.val[2]), q4b.val[3], q8b.val[3]); + + int ls1 = ((x[ibl].scales_l[ib] & 0xf) | ((h << 4) & 0x30)) - 32; + int ls2 = ((x[ibl].scales_l[ib] >> 4) | ((h << 2) & 0x30)) - 32; + h >>= 4; + sumi1 += vaddvq_s32(prod_1) * ls1; + sumi2 += vaddvq_s32(prod_2) * ls2; } - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + sumf += GGML_FP16_TO_FP32(x[ibl].d) * y[ibl].d * (sumi1 + sumi2); } - *s = hsum_float_8(acc); + *s = sumf; -#elif defined __AVX__ +#elif defined __AVX2__ - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i m3 = _mm_set1_epi8(3); - const __m128i m32s = _mm_set1_epi8(32); - const __m128i m2 = _mm_set1_epi8(2); + const __m128i values128 = _mm_loadu_si128((const __m128i*)kvalues_iq4nl); + const __m128i m4b = _mm_set1_epi8(0x0f); + + __m256 accum = _mm256_setzero_ps(); + for (int ibl = 0; ibl < nb; ++ibl) { + const uint8_t * qs = x[ibl].qs; + const int8_t * q8 = y[ibl].qs; + uint16_t sh = x[ibl].scales_h; + __m256i sumi1 = _mm256_setzero_si256(); + __m256i sumi2 = _mm256_setzero_si256(); + for (int ib = 0; ib < QK_K/32; ib += 2) { + const __m128i q4bits_1 = _mm_loadu_si128((const __m128i*)qs); qs += 16; + const __m128i q4bits_2 = _mm_loadu_si128((const __m128i*)qs); qs += 16; + const __m256i q8b_1 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q8b_2 = _mm256_loadu_si256((const __m256i *)q8); q8 += 32; + const __m256i q4b_1 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_1, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_1, m4b))); + const __m256i q4b_2 = MM256_SET_M128I(_mm_shuffle_epi8(values128, _mm_and_si128(_mm_srli_epi16(q4bits_2, 4), m4b)), + _mm_shuffle_epi8(values128, _mm_and_si128(q4bits_2, m4b))); + const __m256i p16_1 = mul_add_epi8(q4b_1, q8b_1); + const __m256i p16_2 = mul_add_epi8(q4b_2, q8b_2); + const int16_t ls1 = ((x[ibl].scales_l[ib/2] & 0xf) | ((sh << 4) & 0x30)) - 32; + const int16_t ls2 = ((x[ibl].scales_l[ib/2] >> 4) | ((sh << 2) & 0x30)) - 32; + sh >>= 4; + const __m256i p_1 = _mm256_madd_epi16(p16_1, _mm256_set1_epi16(ls1)); + const __m256i p_2 = _mm256_madd_epi16(p16_2, _mm256_set1_epi16(ls2)); + sumi1 = _mm256_add_epi32(p_1, sumi1); + sumi2 = _mm256_add_epi32(p_2, sumi2); + } + accum = _mm256_fmadd_ps(_mm256_set1_ps(GGML_FP16_TO_FP32(x[ibl].d)*y[ibl].d), + _mm256_cvtepi32_ps(_mm256_add_epi32(sumi1, sumi2)), accum); + } + + *s = hsum_float_8(accum); - __m256 acc = _mm256_setzero_ps(); +#else + float sumf = 0; + for (int ibl = 0; ibl < nb; ++ibl) { + const float d4d8 = GGML_FP16_TO_FP32(x[ibl].d) * y[ibl].d; + uint16_t h = x[ibl].scales_h; + const uint8_t * qs = x[ibl].qs; + const int8_t * q8 = y[ibl].qs; + for (int ib = 0; ib < QK_K/32; ib += 2) { + const uint8_t ls1 = (x[ibl].scales_l[ib/2] & 0xf) | ((h << 4) & 0x30); + const uint8_t ls2 = (x[ibl].scales_l[ib/2] >> 4) | ((h << 2) & 0x30); + h >>= 4; + const float d1 = d4d8*(ls1 - 32); + const float d2 = d4d8*(ls2 - 32); + int sumi1 = 0, sumi2 = 0; + for (int j = 0; j < 16; ++j) { + sumi1 += q8[j+ 0] * kvalues_iq4nl[qs[j] & 0xf]; + sumi2 += q8[j+16] * kvalues_iq4nl[qs[j] >> 4]; + } + sumf += d1 * (sumi1 + sumi2); + qs += 16; + q8 += 32; + sumi1 = sumi2 = 0; + for (int j = 0; j < 16; ++j) { + sumi1 += q8[j+ 0] * kvalues_iq4nl[qs[j] & 0xf]; + sumi2 += q8[j+16] * kvalues_iq4nl[qs[j] >> 4]; + } + sumf += d2 * (sumi1 + sumi2); + qs += 16; + q8 += 32; + } + } + *s = sumf; +#endif +#endif +} - for (int i = 0; i < nb; ++i) { +// ================================ IQ2 quantization ============================================= + +typedef struct { + uint64_t * grid; + int * map; + uint16_t * neighbours; +} iq2_entry_t; + +static iq2_entry_t iq2_data[4] = { + {NULL, NULL, NULL}, + {NULL, NULL, NULL}, + {NULL, NULL, NULL}, + {NULL, NULL, NULL}, +}; + +static inline int iq2_data_index(enum ggml_type type) { + GGML_ASSERT(type == GGML_TYPE_IQ2_XXS || type == GGML_TYPE_IQ2_XS || type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M || type == GGML_TYPE_IQ2_S); + return type == GGML_TYPE_IQ2_XXS ? 0 : + type == GGML_TYPE_IQ2_XS ? 1 : + type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M ? 2 : 3; +} - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); +static inline int iq2_grid_size(enum ggml_type type) { + GGML_ASSERT(type == GGML_TYPE_IQ2_XXS || type == GGML_TYPE_IQ2_XS || type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M || type == GGML_TYPE_IQ2_S); + return type == GGML_TYPE_IQ2_XXS ? 256 : + type == GGML_TYPE_IQ2_XS ? 512 : + type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M ? NGRID_IQ1S : 1024; +} - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; +static int iq2_compare_func(const void * left, const void * right) { + const int * l = (const int *)left; + const int * r = (const int *)right; + return l[0] < r[0] ? -1 : l[0] > r[0] ? 1 : l[1] < r[1] ? -1 : l[1] > r[1] ? 1 : 0; +} - const __m128i scales = _mm_loadu_si128((const __m128i*)x[i].scales); +void iq2xs_init_impl(enum ggml_type type) { + const int gindex = iq2_data_index(type); + const int grid_size = iq2_grid_size(type); + if (iq2_data[gindex].grid) { + return; + } + static const uint16_t kgrid_2bit_256[256] = { + 0, 2, 5, 8, 10, 17, 20, 32, 34, 40, 42, 65, 68, 80, 88, 97, + 100, 128, 130, 138, 162, 257, 260, 272, 277, 320, 388, 408, 512, 514, 546, 642, + 1025, 1028, 1040, 1057, 1060, 1088, 1090, 1096, 1120, 1153, 1156, 1168, 1188, 1280, 1282, 1288, + 1312, 1350, 1385, 1408, 1425, 1545, 1552, 1600, 1668, 1700, 2048, 2053, 2056, 2068, 2088, 2113, + 2116, 2128, 2130, 2184, 2308, 2368, 2562, 2580, 4097, 4100, 4112, 4129, 4160, 4192, 4228, 4240, + 4245, 4352, 4360, 4384, 4432, 4442, 4480, 4644, 4677, 5120, 5128, 5152, 5157, 5193, 5248, 5400, + 5474, 5632, 5654, 6145, 6148, 6160, 6208, 6273, 6400, 6405, 6560, 6737, 8192, 8194, 8202, 8260, + 8289, 8320, 8322, 8489, 8520, 8704, 8706, 9217, 9220, 9232, 9280, 9302, 9472, 9537, 9572, 9872, + 10248, 10272, 10388, 10820, 16385, 16388, 16400, 16408, 16417, 16420, 16448, 16456, 16470, 16480, 16513, 16516, + 16528, 16640, 16672, 16737, 16768, 16773, 16897, 16912, 16968, 16982, 17000, 17408, 17416, 17440, 17536, 17561, + 17682, 17700, 17920, 18433, 18436, 18448, 18496, 18501, 18688, 18776, 18785, 18818, 19013, 19088, 20480, 20488, + 20497, 20505, 20512, 20608, 20616, 20740, 20802, 20900, 21137, 21648, 21650, 21770, 22017, 22100, 22528, 22545, + 22553, 22628, 22848, 23048, 24580, 24592, 24640, 24680, 24832, 24917, 25112, 25184, 25600, 25605, 25872, 25874, + 25988, 26690, 32768, 32770, 32778, 32833, 32898, 33028, 33048, 33088, 33297, 33793, 33796, 33808, 33813, 33856, + 33888, 34048, 34118, 34196, 34313, 34368, 34400, 34818, 35076, 35345, 36868, 36880, 36900, 36928, 37025, 37142, + 37248, 37445, 37888, 37922, 37956, 38225, 39041, 39200, 40962, 41040, 41093, 41225, 41472, 42008, 43088, 43268, + }; + static const uint16_t kgrid_2bit_512[512] = { + 0, 2, 5, 8, 10, 17, 20, 22, 25, 32, 34, 37, 40, 65, 68, 70, + 73, 80, 82, 85, 88, 97, 100, 128, 130, 133, 136, 145, 148, 153, 160, 257, + 260, 262, 265, 272, 274, 277, 280, 282, 289, 292, 320, 322, 325, 328, 337, 340, + 352, 360, 385, 388, 400, 512, 514, 517, 520, 529, 532, 544, 577, 580, 592, 597, + 640, 650, 1025, 1028, 1030, 1033, 1040, 1042, 1045, 1048, 1057, 1060, 1088, 1090, 1093, 1096, + 1105, 1108, 1110, 1120, 1153, 1156, 1168, 1280, 1282, 1285, 1288, 1297, 1300, 1312, 1345, 1348, + 1360, 1377, 1408, 1537, 1540, 1552, 1574, 1600, 1602, 1668, 2048, 2050, 2053, 2056, 2058, 2065, + 2068, 2080, 2085, 2113, 2116, 2128, 2136, 2176, 2208, 2218, 2305, 2308, 2320, 2368, 2433, 2441, + 2560, 2592, 2600, 2710, 2720, 4097, 4100, 4102, 4105, 4112, 4114, 4117, 4120, 4129, 4132, 4160, + 4162, 4165, 4168, 4177, 4180, 4192, 4202, 4225, 4228, 4240, 4352, 4354, 4357, 4360, 4369, 4372, + 4384, 4417, 4420, 4432, 4480, 4500, 4502, 4609, 4612, 4614, 4624, 4672, 4704, 5120, 5122, 5125, + 5128, 5137, 5140, 5152, 5185, 5188, 5193, 5200, 5220, 5248, 5377, 5380, 5392, 5440, 5632, 5652, + 5705, 6145, 6148, 6160, 6162, 6208, 6228, 6278, 6400, 6405, 6502, 6737, 6825, 8192, 8194, 8197, + 8200, 8202, 8209, 8212, 8224, 8257, 8260, 8272, 8320, 8352, 8449, 8452, 8464, 8512, 8520, 8549, + 8704, 8738, 8832, 8872, 9217, 9220, 9232, 9257, 9280, 9472, 9537, 9554, 9625, 9729, 9754, 9894, + 10240, 10248, 10250, 10272, 10325, 10376, 10402, 10600, 10640, 10760, 10784, 10882, 10888, 10890, 16385, 16388, + 16390, 16393, 16400, 16402, 16405, 16408, 16417, 16420, 16448, 16450, 16453, 16456, 16458, 16465, 16468, 16480, + 16485, 16513, 16516, 16528, 16640, 16642, 16645, 16648, 16657, 16660, 16672, 16705, 16708, 16720, 16768, 16773, + 16802, 16897, 16900, 16912, 16914, 16937, 16960, 17408, 17410, 17413, 17416, 17425, 17428, 17433, 17440, 17473, + 17476, 17488, 17536, 17556, 17665, 17668, 17680, 17700, 17728, 17818, 17920, 17930, 17988, 18000, 18433, 18436, + 18448, 18496, 18501, 18516, 18530, 18688, 18705, 18756, 18768, 18793, 18948, 20480, 20482, 20485, 20488, 20497, + 20500, 20512, 20520, 20545, 20548, 20560, 20608, 20737, 20740, 20752, 20757, 20800, 20802, 20992, 21060, 21162, + 21505, 21508, 21520, 21537, 21568, 21600, 21633, 21665, 21760, 21768, 21888, 21896, 22049, 22120, 22177, 22528, + 22548, 22593, 22608, 22681, 22810, 22848, 22850, 23173, 24577, 24580, 24592, 24640, 24660, 24674, 24710, 24745, + 24832, 25124, 25162, 25234, 25600, 25622, 25872, 25920, 25925, 26020, 26625, 26730, 26917, 27142, 27220, 27234, + 32768, 32770, 32773, 32776, 32785, 32788, 32800, 32810, 32833, 32836, 32848, 32896, 32898, 32936, 32938, 33025, + 33028, 33030, 33040, 33088, 33105, 33113, 33280, 33312, 33408, 33410, 33440, 33448, 33793, 33796, 33808, 33810, + 33813, 33856, 33888, 33929, 34048, 34116, 34213, 34328, 34410, 34816, 34824, 34853, 34906, 34944, 34946, 34984, + 35078, 35362, 35456, 35464, 35478, 35496, 36865, 36868, 36880, 36928, 36950, 36996, 37120, 37154, 37220, 37462, + 37513, 37888, 37893, 37956, 37968, 37976, 38185, 38288, 38290, 38465, 38993, 39078, 39241, 39445, 39520, 40960, + 40962, 40968, 40970, 40992, 41002, 41120, 41297, 41305, 41382, 41472, 41474, 41480, 41514, 41600, 41632, 42048, + 42133, 42597, 42648, 43018, 43040, 43042, 43048, 43168, 43176, 43268, 43396, 43398, 43560, 43562, 43665, 43690, + }; + static const uint16_t kgrid_1bit_2048[NGRID_IQ1S] = { + 0, 2, 5, 8, 10, 17, 21, 32, 34, 40, 42, 69, 81, 84, 86, 101, + 128, 130, 136, 138, 149, 160, 162, 168, 170, 260, 261, 273, 276, 278, 281, 282, + 293, 321, 326, 329, 338, 341, 346, 353, 356, 358, 360, 389, 401, 404, 406, 421, + 512, 514, 520, 522, 533, 544, 546, 552, 554, 581, 593, 601, 612, 617, 640, 642, + 648, 650, 657, 661, 665, 672, 674, 680, 682, 1041, 1044, 1046, 1061, 1089, 1097, 1109, + 1114, 1124, 1125, 1169, 1177, 1189, 1281, 1284, 1285, 1286, 1301, 1304, 1306, 1321, 1344, 1349, + 1354, 1360, 1361, 1364, 1365, 1366, 1369, 1376, 1378, 1381, 1384, 1386, 1409, 1425, 1429, 1432, + 1434, 1441, 1444, 1445, 1446, 1449, 1556, 1561, 1601, 1604, 1616, 1618, 1621, 1624, 1632, 1633, + 1638, 1641, 1669, 1681, 1684, 1689, 2048, 2050, 2056, 2058, 2069, 2080, 2082, 2088, 2090, 2117, + 2129, 2134, 2149, 2176, 2178, 2184, 2186, 2197, 2208, 2210, 2216, 2218, 2309, 2321, 2324, 2329, + 2340, 2341, 2369, 2384, 2385, 2389, 2401, 2404, 2409, 2449, 2452, 2454, 2457, 2469, 2560, 2562, + 2568, 2570, 2581, 2592, 2594, 2600, 2602, 2629, 2641, 2649, 2657, 2661, 2688, 2690, 2693, 2696, + 2698, 2709, 2720, 2722, 2728, 2730, 4112, 4113, 4116, 4121, 4132, 4133, 4161, 4164, 4176, 4181, + 4184, 4193, 4196, 4197, 4201, 4241, 4244, 4246, 4257, 4261, 4353, 4356, 4358, 4361, 4368, 4370, + 4373, 4376, 4385, 4388, 4393, 4421, 4426, 4432, 4433, 4434, 4436, 4437, 4438, 4441, 4448, 4453, + 4484, 4498, 4501, 4513, 4516, 4625, 4628, 4630, 4645, 4672, 4678, 4681, 4690, 4693, 4696, 4698, + 4708, 4710, 4741, 4753, 4756, 4758, 4773, 5121, 5126, 5129, 5140, 5141, 5144, 5145, 5153, 5158, + 5185, 5189, 5190, 5192, 5194, 5201, 5204, 5205, 5206, 5209, 5218, 5221, 5224, 5252, 5257, 5264, + 5268, 5269, 5272, 5273, 5274, 5281, 5284, 5285, 5289, 5378, 5381, 5386, 5393, 5396, 5397, 5398, + 5401, 5408, 5410, 5413, 5416, 5418, 5441, 5444, 5445, 5446, 5457, 5458, 5460, 5461, 5462, 5465, + 5466, 5473, 5476, 5477, 5478, 5481, 5504, 5506, 5508, 5509, 5512, 5514, 5520, 5521, 5524, 5525, + 5526, 5529, 5530, 5536, 5538, 5541, 5633, 5636, 5637, 5638, 5653, 5654, 5656, 5658, 5665, 5670, + 5696, 5698, 5700, 5701, 5704, 5706, 5713, 5717, 5718, 5720, 5721, 5729, 5732, 5733, 5736, 5737, + 5738, 5766, 5770, 5778, 5781, 5796, 5801, 6161, 6166, 6181, 6209, 6212, 6214, 6217, 6224, 6229, + 6232, 6234, 6240, 6241, 6244, 6246, 6249, 6277, 6289, 6292, 6309, 6416, 6418, 6421, 6426, 6433, + 6437, 6466, 6468, 6469, 6472, 6481, 6484, 6485, 6486, 6489, 6490, 6496, 6501, 6506, 6537, 6545, + 6546, 6549, 6552, 6561, 6566, 6569, 6665, 6678, 6692, 6694, 6724, 6726, 6729, 6736, 6738, 6741, + 6744, 6753, 6758, 6761, 6789, 6801, 6806, 6810, 8192, 8194, 8200, 8202, 8213, 8224, 8226, 8229, + 8232, 8234, 8261, 8273, 8281, 8289, 8293, 8320, 8322, 8328, 8330, 8341, 8352, 8354, 8357, 8360, + 8362, 8453, 8465, 8468, 8473, 8485, 8514, 8516, 8521, 8533, 8536, 8538, 8545, 8548, 8549, 8550, + 8581, 8592, 8598, 8601, 8613, 8705, 8712, 8714, 8721, 8725, 8736, 8738, 8744, 8746, 8773, 8785, + 8790, 8793, 8805, 8833, 8840, 8842, 8849, 8853, 8864, 8866, 8872, 8874, 9221, 9236, 9238, 9241, + 9253, 9284, 9285, 9286, 9289, 9298, 9301, 9304, 9306, 9318, 9349, 9361, 9364, 9369, 9377, 9381, + 9481, 9493, 9505, 9513, 9536, 9541, 9544, 9553, 9556, 9557, 9561, 9570, 9573, 9576, 9609, 9616, + 9620, 9621, 9624, 9626, 9633, 9636, 9638, 9641, 9733, 9744, 9746, 9753, 9765, 9793, 9801, 9813, + 9824, 9825, 9833, 9860, 9862, 9872, 9882, 10240, 10242, 10248, 10250, 10261, 10272, 10274, 10280, 10282, + 10309, 10321, 10324, 10341, 10368, 10370, 10376, 10378, 10400, 10402, 10408, 10410, 10505, 10513, 10516, 10521, + 10533, 10566, 10569, 10578, 10581, 10593, 10596, 10598, 10601, 10629, 10640, 10646, 10649, 10660, 10661, 10752, + 10754, 10760, 10762, 10784, 10786, 10792, 10794, 10821, 10833, 10838, 10841, 10853, 10880, 10882, 10888, 10890, + 10901, 10912, 10914, 10920, 10922, 16389, 16401, 16406, 16421, 16457, 16466, 16469, 16472, 16474, 16481, 16484, + 16486, 16532, 16537, 16545, 16550, 16640, 16641, 16644, 16646, 16649, 16658, 16661, 16662, 16664, 16666, 16673, + 16678, 16681, 16709, 16712, 16714, 16721, 16724, 16725, 16726, 16729, 16730, 16741, 16744, 16746, 16769, 16772, + 16774, 16784, 16786, 16789, 16800, 16801, 16802, 16901, 16913, 16916, 16918, 16933, 16961, 16978, 16981, 16986, + 16996, 17001, 17033, 17044, 17061, 17409, 17429, 17433, 17449, 17477, 17480, 17482, 17489, 17492, 17493, 17494, + 17505, 17506, 17509, 17512, 17514, 17537, 17542, 17545, 17552, 17554, 17557, 17568, 17569, 17577, 17665, 17666, + 17669, 17674, 17681, 17684, 17685, 17686, 17689, 17696, 17701, 17706, 17729, 17732, 17733, 17734, 17737, 17744, + 17745, 17748, 17749, 17750, 17752, 17753, 17761, 17764, 17765, 17766, 17769, 17794, 17796, 17797, 17800, 17809, + 17812, 17813, 17814, 17817, 17818, 17829, 17832, 17834, 17921, 17925, 17929, 17940, 17941, 17944, 17946, 17953, + 17956, 17961, 17984, 17986, 17989, 17992, 18000, 18001, 18002, 18005, 18006, 18009, 18018, 18021, 18024, 18049, + 18053, 18058, 18068, 18069, 18081, 18084, 18086, 18437, 18449, 18453, 18458, 18469, 18498, 18505, 18512, 18517, + 18520, 18529, 18532, 18534, 18537, 18565, 18577, 18580, 18582, 18585, 18597, 18689, 18693, 18694, 18698, 18704, + 18708, 18709, 18712, 18721, 18724, 18726, 18752, 18757, 18762, 18769, 18770, 18772, 18773, 18774, 18777, 18784, + 18786, 18789, 18790, 18794, 18822, 18825, 18834, 18837, 18838, 18840, 18849, 18852, 18854, 18857, 18966, 19012, + 19014, 19017, 19029, 19032, 19034, 19044, 19049, 19092, 19109, 20481, 20484, 20485, 20486, 20489, 20498, 20501, + 20506, 20513, 20516, 20521, 20544, 20549, 20552, 20561, 20564, 20565, 20566, 20569, 20581, 20584, 20614, 20617, + 20629, 20632, 20640, 20641, 20646, 20649, 20741, 20744, 20745, 20746, 20753, 20756, 20757, 20758, 20760, 20761, + 20768, 20773, 20774, 20776, 20778, 20801, 20804, 20805, 20806, 20809, 20816, 20817, 20818, 20820, 20821, 20822, + 20824, 20825, 20826, 20833, 20836, 20837, 20838, 20841, 20866, 20869, 20881, 20884, 20885, 20886, 20889, 20896, + 20901, 20906, 20993, 20998, 21010, 21013, 21018, 21025, 21028, 21058, 21061, 21066, 21073, 21076, 21077, 21078, + 21081, 21090, 21093, 21125, 21136, 21138, 21141, 21145, 21146, 21156, 21508, 21509, 21521, 21524, 21525, 21526, + 21528, 21529, 21537, 21541, 21544, 21546, 21569, 21572, 21573, 21574, 21577, 21578, 21584, 21585, 21588, 21589, + 21590, 21592, 21593, 21594, 21601, 21602, 21604, 21605, 21606, 21609, 21632, 21640, 21642, 21649, 21652, 21653, + 21654, 21657, 21665, 21668, 21669, 21674, 21761, 21762, 21764, 21765, 21766, 21769, 21776, 21777, 21778, 21780, + 21781, 21782, 21785, 21786, 21793, 21796, 21797, 21798, 21801, 21824, 21825, 21826, 21828, 21829, 21830, 21832, + 21833, 21840, 21841, 21842, 21844, 21845, 21846, 21848, 21849, 21850, 21856, 21857, 21860, 21861, 21862, 21864, + 21865, 21866, 21889, 21892, 21893, 21897, 21898, 21904, 21905, 21908, 21909, 21910, 21912, 21913, 21921, 21924, + 21925, 21926, 21929, 22016, 22017, 22018, 22020, 22022, 22024, 22025, 22033, 22036, 22037, 22040, 22041, 22048, + 22049, 22050, 22052, 22053, 22054, 22056, 22057, 22081, 22085, 22086, 22088, 22089, 22090, 22096, 22097, 22098, + 22100, 22101, 22102, 22104, 22105, 22106, 22113, 22116, 22117, 22121, 22146, 22149, 22150, 22152, 22153, 22154, + 22161, 22165, 22170, 22178, 22181, 22182, 22184, 22185, 22532, 22533, 22534, 22537, 22544, 22549, 22552, 22561, + 22570, 22597, 22600, 22602, 22609, 22612, 22613, 22614, 22616, 22617, 22624, 22626, 22628, 22629, 22658, 22665, + 22672, 22674, 22677, 22680, 22689, 22697, 22785, 22786, 22789, 22794, 22801, 22804, 22805, 22806, 22809, 22821, + 22849, 22852, 22853, 22854, 22857, 22864, 22865, 22866, 22868, 22869, 22870, 22872, 22873, 22874, 22881, 22884, + 22885, 22886, 22889, 22913, 22917, 22921, 22929, 22932, 22933, 22934, 22936, 22937, 22949, 23044, 23048, 23061, + 23066, 23072, 23077, 23078, 23081, 23109, 23112, 23113, 23121, 23125, 23126, 23128, 23129, 23138, 23141, 23144, + 23146, 23169, 23178, 23186, 23189, 23190, 23192, 23194, 23201, 24581, 24596, 24598, 24601, 24613, 24644, 24656, + 24661, 24662, 24664, 24666, 24673, 24676, 24678, 24681, 24705, 24726, 24741, 24833, 24836, 24838, 24841, 24850, + 24853, 24865, 24866, 24870, 24873, 24901, 24905, 24913, 24917, 24918, 24921, 24933, 24934, 24938, 24964, 24970, + 24978, 24981, 24993, 24998, 25001, 25105, 25110, 25113, 25152, 25153, 25158, 25173, 25174, 25176, 25184, 25221, + 25233, 25238, 25253, 25617, 25618, 25621, 25622, 25626, 25633, 25638, 25641, 25664, 25666, 25669, 25672, 25674, + 25681, 25684, 25685, 25686, 25689, 25690, 25696, 25698, 25701, 25732, 25733, 25737, 25744, 25746, 25748, 25749, + 25750, 25752, 25754, 25761, 25764, 25769, 25861, 25864, 25866, 25873, 25877, 25878, 25881, 25924, 25925, 25926, + 25929, 25936, 25937, 25940, 25941, 25942, 25945, 25953, 25956, 25957, 25958, 25961, 25990, 25993, 25994, 26001, + 26005, 26006, 26009, 26010, 26018, 26021, 26022, 26024, 26114, 26121, 26133, 26144, 26150, 26152, 26153, 26176, + 26181, 26184, 26186, 26193, 26196, 26197, 26198, 26200, 26202, 26208, 26213, 26216, 26240, 26242, 26245, 26250, + 26260, 26262, 26264, 26265, 26272, 26276, 26278, 26282, 26646, 26649, 26661, 26689, 26706, 26709, 26714, 26721, + 26729, 26757, 26769, 26776, 26790, 26881, 26884, 26896, 26901, 26913, 26916, 26918, 26921, 26944, 26945, 26949, + 26950, 26952, 26961, 26964, 26965, 26966, 26969, 26976, 26981, 26986, 27010, 27012, 27018, 27029, 27041, 27044, + 27045, 27049, 27153, 27158, 27160, 27201, 27204, 27209, 27216, 27221, 27224, 27226, 27236, 27237, 27241, 27270, + 27284, 27288, 27290, 27302, 32768, 32770, 32776, 32778, 32800, 32802, 32808, 32810, 32837, 32848, 32849, 32852, + 32854, 32857, 32869, 32896, 32898, 32904, 32906, 32917, 32928, 32930, 32936, 32938, 33029, 33041, 33044, 33046, + 33049, 33061, 33089, 33092, 33097, 33104, 33106, 33109, 33110, 33112, 33113, 33124, 33126, 33129, 33157, 33161, + 33172, 33174, 33177, 33189, 33280, 33282, 33288, 33290, 33301, 33312, 33314, 33320, 33322, 33361, 33364, 33369, + 33381, 33408, 33410, 33416, 33418, 33429, 33440, 33442, 33448, 33450, 33812, 33817, 33857, 33860, 33873, 33877, + 33882, 33889, 33892, 33897, 33940, 33945, 34049, 34057, 34066, 34069, 34074, 34086, 34089, 34112, 34113, 34117, + 34120, 34129, 34132, 34133, 34134, 34137, 34138, 34149, 34150, 34152, 34154, 34177, 34180, 34182, 34185, 34192, + 34194, 34197, 34200, 34214, 34321, 34326, 34329, 34341, 34369, 34372, 34377, 34378, 34384, 34389, 34393, 34394, + 34401, 34406, 34410, 34437, 34449, 34458, 34468, 34816, 34818, 34824, 34826, 34837, 34848, 34850, 34856, 34858, + 34881, 34885, 34897, 34900, 34905, 34917, 34921, 34944, 34946, 34952, 34954, 34965, 34976, 34978, 34984, 34986, + 35077, 35078, 35089, 35092, 35094, 35109, 35137, 35140, 35142, 35145, 35152, 35154, 35157, 35162, 35169, 35172, + 35205, 35222, 35225, 35237, 35328, 35330, 35336, 35338, 35349, 35360, 35362, 35368, 35370, 35397, 35409, 35412, + 35414, 35456, 35458, 35464, 35466, 35477, 35488, 35490, 35496, 35498, 36869, 36881, 36886, 36888, 36889, 36901, + 36929, 36934, 36937, 36949, 36952, 36954, 36969, 36970, 36997, 37009, 37012, 37014, 37017, 37029, 37121, 37124, + 37126, 37129, 37136, 37141, 37144, 37146, 37153, 37156, 37158, 37161, 37184, 37189, 37200, 37201, 37204, 37205, + 37206, 37209, 37218, 37221, 37252, 37254, 37266, 37269, 37272, 37281, 37284, 37286, 37289, 37381, 37393, 37396, + 37401, 37413, 37444, 37446, 37449, 37456, 37458, 37461, 37464, 37478, 37481, 37509, 37524, 37526, 37545, 37889, + 37892, 37894, 37904, 37909, 37912, 37926, 37952, 37962, 37969, 37972, 37973, 37974, 37976, 37977, 37984, 37985, + 37986, 37989, 38020, 38022, 38034, 38036, 38037, 38040, 38049, 38057, 38144, 38149, 38152, 38154, 38160, 38161, + 38164, 38165, 38166, 38169, 38177, 38181, 38185, 38186, 38209, 38212, 38213, 38214, 38217, 38224, 38225, 38226, + 38228, 38229, 38230, 38232, 38233, 38234, 38241, 38244, 38245, 38246, 38249, 38273, 38277, 38280, 38289, 38290, + 38292, 38293, 38294, 38297, 38298, 38304, 38306, 38309, 38312, 38314, 38401, 38404, 38416, 38421, 38425, 38432, + 38438, 38441, 38469, 38472, 38473, 38481, 38482, 38485, 38486, 38489, 38501, 38504, 38530, 38532, 38537, 38538, + 38546, 38548, 38549, 38564, 38566, 38569, 38917, 38934, 38937, 38949, 38977, 38982, 38992, 38994, 38997, 38998, + 39002, 39012, 39013, 39045, 39057, 39062, 39065, 39077, 39172, 39174, 39177, 39184, 39186, 39189, 39192, 39194, + 39200, 39201, 39204, 39206, 39232, 39234, 39237, 39240, 39242, 39249, 39252, 39253, 39254, 39257, 39266, 39269, + 39270, 39274, 39297, 39300, 39312, 39314, 39317, 39322, 39329, 39334, 39429, 39445, 39461, 39492, 39494, 39497, + 39504, 39509, 39512, 39521, 39557, 39569, 39572, 39573, 39574, 40960, 40962, 40968, 40970, 40981, 40992, 40994, + 41000, 41002, 41029, 41041, 41044, 41046, 41049, 41088, 41090, 41096, 41098, 41109, 41120, 41122, 41128, 41130, + 41221, 41225, 41233, 41236, 41238, 41241, 41242, 41286, 41289, 41297, 41301, 41304, 41306, 41313, 41316, 41349, + 41360, 41362, 41366, 41369, 41474, 41480, 41482, 41488, 41497, 41506, 41512, 41514, 41541, 41553, 41558, 41561, + 41573, 41600, 41602, 41608, 41610, 41621, 41632, 41634, 41640, 41642, 42009, 42021, 42049, 42052, 42064, 42068, + 42069, 42072, 42074, 42081, 42085, 42086, 42088, 42089, 42117, 42246, 42249, 42256, 42258, 42261, 42264, 42278, + 42281, 42306, 42309, 42321, 42324, 42325, 42326, 42329, 42341, 42346, 42369, 42372, 42373, 42374, 42377, 42386, + 42389, 42392, 42501, 42513, 42518, 42522, 42529, 42533, 42564, 42566, 42570, 42578, 42581, 42582, 42584, 42592, + 42594, 42630, 42640, 42645, 42646, 42649, 42657, 42660, 42662, 43008, 43010, 43016, 43018, 43040, 43042, 43048, + 43050, 43089, 43092, 43094, 43097, 43136, 43138, 43144, 43146, 43157, 43168, 43170, 43176, 43178, 43269, 43284, + 43289, 43297, 43301, 43329, 43344, 43349, 43354, 43361, 43366, 43369, 43408, 43414, 43520, 43522, 43528, 43530, + 43552, 43554, 43560, 43562, 43601, 43604, 43606, 43648, 43650, 43656, 43658, 43669, 43680, 43682, 43688, 43690, + }; + static const uint16_t kgrid_2bit_1024[1024] = { + 0, 2, 5, 8, 10, 17, 20, 22, 25, 32, 34, 37, 40, 65, 68, 70, + 73, 80, 82, 85, 88, 97, 100, 102, 105, 128, 130, 133, 136, 145, 148, 160, + 165, 170, 257, 260, 262, 265, 272, 274, 277, 280, 289, 292, 320, 322, 325, 328, + 337, 340, 342, 345, 352, 357, 360, 385, 388, 400, 402, 405, 417, 420, 512, 514, + 517, 520, 529, 532, 544, 554, 577, 580, 582, 585, 592, 597, 640, 645, 650, 660, + 674, 1025, 1028, 1030, 1033, 1040, 1042, 1045, 1048, 1057, 1060, 1062, 1065, 1088, 1090, 1093, + 1096, 1098, 1105, 1108, 1110, 1113, 1120, 1122, 1125, 1153, 1156, 1158, 1161, 1168, 1173, 1176, + 1185, 1188, 1280, 1282, 1285, 1288, 1290, 1297, 1300, 1302, 1305, 1312, 1317, 1320, 1345, 1348, + 1350, 1353, 1360, 1362, 1365, 1368, 1377, 1380, 1408, 1410, 1413, 1416, 1425, 1428, 1440, 1537, + 1540, 1542, 1545, 1552, 1557, 1600, 1605, 1608, 1617, 1620, 1632, 1665, 1668, 1680, 2048, 2050, + 2053, 2056, 2065, 2068, 2070, 2073, 2080, 2085, 2090, 2113, 2116, 2118, 2121, 2128, 2130, 2133, + 2136, 2145, 2148, 2176, 2181, 2196, 2218, 2305, 2308, 2320, 2322, 2325, 2328, 2337, 2368, 2373, + 2376, 2385, 2388, 2400, 2433, 2448, 2560, 2577, 2580, 2594, 2600, 2602, 2640, 2713, 4097, 4100, + 4102, 4105, 4112, 4114, 4117, 4120, 4129, 4132, 4134, 4160, 4162, 4165, 4168, 4177, 4180, 4182, + 4185, 4192, 4194, 4197, 4200, 4225, 4228, 4230, 4240, 4245, 4248, 4257, 4260, 4352, 4354, 4357, + 4360, 4362, 4369, 4372, 4374, 4377, 4384, 4386, 4389, 4392, 4417, 4420, 4422, 4425, 4432, 4434, + 4437, 4440, 4449, 4452, 4480, 4482, 4485, 4488, 4497, 4500, 4609, 4612, 4617, 4624, 4629, 4641, + 4644, 4672, 4677, 4689, 4692, 4737, 4740, 4752, 5120, 5122, 5125, 5128, 5137, 5140, 5142, 5145, + 5152, 5157, 5160, 5185, 5188, 5190, 5193, 5200, 5202, 5205, 5208, 5217, 5220, 5248, 5250, 5253, + 5256, 5265, 5268, 5280, 5377, 5380, 5382, 5385, 5392, 5394, 5397, 5400, 5409, 5412, 5440, 5442, + 5445, 5448, 5457, 5460, 5472, 5505, 5508, 5520, 5632, 5637, 5640, 5649, 5652, 5664, 5697, 5700, + 5712, 5760, 5802, 6145, 6148, 6150, 6153, 6160, 6165, 6168, 6177, 6208, 6210, 6213, 6216, 6225, + 6228, 6240, 6273, 6276, 6400, 6402, 6405, 6408, 6417, 6420, 6432, 6465, 6468, 6480, 6505, 6562, + 6660, 6672, 6720, 6742, 8192, 8194, 8197, 8200, 8209, 8212, 8214, 8217, 8224, 8229, 8234, 8257, + 8260, 8272, 8274, 8277, 8292, 8320, 8330, 8340, 8362, 8449, 8452, 8464, 8466, 8469, 8481, 8512, + 8514, 8517, 8529, 8532, 8544, 8577, 8580, 8592, 8704, 8714, 8738, 8744, 8746, 8772, 8784, 8840, + 8842, 8872, 9217, 9220, 9222, 9225, 9232, 9237, 9240, 9249, 9252, 9280, 9282, 9285, 9288, 9297, + 9300, 9312, 9345, 9348, 9360, 9472, 9477, 9480, 9489, 9492, 9504, 9537, 9540, 9552, 9574, 9600, + 9729, 9732, 9744, 9792, 9817, 10240, 10245, 10257, 10260, 10305, 10308, 10320, 10378, 10410, 10497, 10500, + 10512, 10645, 10762, 10786, 10852, 10888, 10890, 16385, 16388, 16390, 16393, 16400, 16402, 16405, 16408, 16410, + 16417, 16420, 16422, 16448, 16450, 16453, 16456, 16458, 16465, 16468, 16470, 16473, 16480, 16482, 16485, 16513, + 16516, 16528, 16533, 16536, 16545, 16548, 16640, 16642, 16645, 16648, 16657, 16660, 16662, 16665, 16672, 16674, + 16677, 16705, 16708, 16710, 16713, 16720, 16722, 16725, 16728, 16737, 16740, 16768, 16770, 16773, 16776, 16785, + 16788, 16800, 16897, 16900, 16912, 16914, 16917, 16920, 16932, 16960, 16965, 16968, 16977, 16980, 16992, 17025, + 17028, 17408, 17410, 17413, 17416, 17418, 17425, 17428, 17430, 17433, 17440, 17442, 17445, 17448, 17473, 17476, + 17478, 17481, 17488, 17490, 17493, 17496, 17505, 17508, 17536, 17538, 17541, 17544, 17553, 17556, 17568, 17665, + 17668, 17670, 17673, 17680, 17682, 17685, 17688, 17697, 17700, 17728, 17730, 17733, 17736, 17745, 17748, 17760, + 17770, 17793, 17796, 17808, 17920, 17922, 17925, 17928, 17937, 17940, 17952, 17985, 17988, 18000, 18048, 18085, + 18433, 18436, 18441, 18448, 18450, 18453, 18456, 18465, 18468, 18496, 18498, 18501, 18504, 18513, 18516, 18528, + 18564, 18576, 18688, 18690, 18693, 18696, 18705, 18708, 18720, 18753, 18756, 18768, 18816, 18838, 18945, 18948, + 18960, 19008, 20480, 20482, 20485, 20488, 20497, 20500, 20502, 20505, 20512, 20514, 20517, 20520, 20545, 20548, + 20550, 20553, 20560, 20562, 20565, 20568, 20577, 20580, 20608, 20610, 20613, 20616, 20625, 20628, 20737, 20740, + 20742, 20745, 20752, 20754, 20757, 20760, 20769, 20772, 20800, 20802, 20805, 20808, 20817, 20820, 20832, 20865, + 20868, 20880, 20992, 20997, 21000, 21009, 21012, 21024, 21057, 21060, 21072, 21097, 21120, 21505, 21508, 21510, + 21513, 21520, 21522, 21525, 21528, 21537, 21540, 21568, 21570, 21573, 21576, 21585, 21588, 21600, 21633, 21636, + 21648, 21760, 21762, 21765, 21768, 21777, 21780, 21792, 21825, 21828, 21840, 21888, 22017, 22020, 22032, 22054, + 22080, 22528, 22530, 22533, 22536, 22545, 22548, 22560, 22593, 22596, 22608, 22618, 22656, 22785, 22788, 22800, + 22848, 23040, 23065, 23173, 23208, 24577, 24580, 24582, 24592, 24594, 24597, 24600, 24609, 24612, 24640, 24645, + 24648, 24657, 24660, 24672, 24708, 24720, 24832, 24834, 24837, 24840, 24849, 24852, 24864, 24897, 24900, 24912, + 24960, 24985, 25092, 25104, 25152, 25174, 25249, 25600, 25605, 25608, 25617, 25620, 25632, 25665, 25668, 25680, + 25728, 25857, 25860, 25872, 25920, 25930, 25960, 26002, 26112, 26260, 26625, 26628, 26640, 26725, 26776, 26880, + 26922, 27202, 27297, 32768, 32770, 32773, 32776, 32785, 32788, 32793, 32800, 32805, 32833, 32836, 32848, 32850, + 32853, 32856, 32865, 32896, 32901, 32913, 32916, 33025, 33028, 33033, 33040, 33042, 33045, 33048, 33057, 33060, + 33088, 33090, 33093, 33096, 33105, 33108, 33153, 33156, 33168, 33193, 33280, 33285, 33290, 33297, 33300, 33345, + 33348, 33360, 33793, 33796, 33798, 33801, 33808, 33810, 33813, 33816, 33825, 33856, 33858, 33861, 33864, 33873, + 33876, 33888, 33921, 33924, 33936, 34048, 34050, 34053, 34056, 34065, 34068, 34080, 34113, 34116, 34128, 34176, + 34186, 34305, 34308, 34320, 34345, 34368, 34816, 34821, 34833, 34836, 34881, 34884, 34896, 34978, 35073, 35076, + 35136, 35173, 35362, 35416, 35418, 35458, 35490, 36865, 36868, 36873, 36880, 36882, 36885, 36888, 36900, 36928, + 36930, 36933, 36936, 36945, 36948, 36960, 36993, 36996, 37008, 37120, 37125, 37137, 37140, 37185, 37188, 37200, + 37210, 37377, 37380, 37392, 37440, 37542, 37888, 37890, 37893, 37896, 37905, 37908, 37920, 37953, 37956, 37968, + 38016, 38038, 38145, 38148, 38160, 38208, 38296, 38305, 38400, 38470, 38500, 38913, 38916, 38928, 38950, 38976, + 39081, 39168, 39241, 39250, 39568, 40960, 40965, 40970, 40980, 40994, 41002, 41025, 41028, 41040, 41122, 41130, + 41280, 41317, 41474, 41482, 41506, 41512, 41514, 41602, 41608, 41610, 41640, 41985, 41988, 42000, 42048, 42121, + 42148, 42240, 42265, 42577, 43018, 43048, 43170, 43348, 43398, 43528, 43530, 43552, 43554, 43560, 43656, 43690, + }; - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); + const int kmap_size = 43692; + //const int nwant = type == GGML_TYPE_IQ1_S ? 3 : 2; + const int nwant = type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M ? 3 : type == GGML_TYPE_IQ2_S ? 1 : 2; + const uint16_t * kgrid = type == GGML_TYPE_IQ2_XXS ? kgrid_2bit_256 : + type == GGML_TYPE_IQ2_XS ? kgrid_2bit_512 : + type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M ? kgrid_1bit_2048 : kgrid_2bit_1024; + uint64_t * kgrid_q2xs; + int * kmap_q2xs; + uint16_t * kneighbors_q2xs; + + //printf("================================================================= %s(grid_size = %d)\n", __func__, grid_size); + uint64_t * the_grid = (uint64_t *)malloc(grid_size*sizeof(uint64_t)); + for (int k = 0; k < grid_size; ++k) { + int8_t * pos = (int8_t *)(the_grid + k); + for (int i = 0; i < 8; ++i) { + int l = (kgrid[k] >> 2*i) & 0x3; + pos[i] = 2*l + 1; + } + } + kgrid_q2xs = the_grid; + iq2_data[gindex].grid = the_grid; + kmap_q2xs = (int *)malloc(kmap_size*sizeof(int)); + iq2_data[gindex].map = kmap_q2xs; + for (int i = 0; i < kmap_size; ++i) kmap_q2xs[i] = -1; + uint64_t aux64; + uint8_t * aux8 = (uint8_t *)&aux64; + for (int i = 0; i < grid_size; ++i) { + aux64 = kgrid_q2xs[i]; + uint16_t index = 0; + for (int k=0; k<8; ++k) { + uint16_t q = (aux8[k] - 1)/2; + index |= (q << 2*k); + } + kmap_q2xs[index] = i; + } + int8_t pos[8]; + int * dist2 = (int *)malloc(2*grid_size*sizeof(int)); + int num_neighbors = 0, num_not_in_map = 0; + for (int i = 0; i < kmap_size; ++i) { + if (kmap_q2xs[i] >= 0) continue; + ++num_not_in_map; + for (int k = 0; k < 8; ++k) { + int l = (i >> 2*k) & 0x3; + pos[k] = 2*l + 1; + } + for (int j = 0; j < grid_size; ++j) { + const int8_t * pg = (const int8_t *)(kgrid_q2xs + j); + int d2 = 0; + for (int k = 0; k < 8; ++k) d2 += (pg[k] - pos[k])*(pg[k] - pos[k]); + dist2[2*j+0] = d2; + dist2[2*j+1] = j; + } + qsort(dist2, grid_size, 2*sizeof(int), iq2_compare_func); + int n = 0; int d2 = dist2[0]; + int nhave = 1; + for (int j = 0; j < grid_size; ++j) { + if (dist2[2*j] > d2) { + if (nhave == nwant) break; + d2 = dist2[2*j]; + ++nhave; + } + ++n; + } + num_neighbors += n; + } + //printf("%s: %d neighbours in total\n", __func__, num_neighbors); + kneighbors_q2xs = (uint16_t *)malloc((num_neighbors + num_not_in_map)*sizeof(uint16_t)); + iq2_data[gindex].neighbours = kneighbors_q2xs; + int counter = 0; + for (int i = 0; i < kmap_size; ++i) { + if (kmap_q2xs[i] >= 0) continue; + for (int k = 0; k < 8; ++k) { + int l = (i >> 2*k) & 0x3; + pos[k] = 2*l + 1; + } + for (int j = 0; j < grid_size; ++j) { + const int8_t * pg = (const int8_t *)(kgrid_q2xs + j); + int d2 = 0; + for (int k = 0; k < 8; ++k) d2 += (pg[k] - pos[k])*(pg[k] - pos[k]); + dist2[2*j+0] = d2; + dist2[2*j+1] = j; + } + qsort(dist2, grid_size, 2*sizeof(int), iq2_compare_func); + kmap_q2xs[i] = -(counter + 1); + int d2 = dist2[0]; + uint16_t * start = &kneighbors_q2xs[counter++]; + int n = 0, nhave = 1; + for (int j = 0; j < grid_size; ++j) { + if (dist2[2*j] > d2) { + if (nhave == nwant) break; + d2 = dist2[2*j]; + ++nhave; + } + kneighbors_q2xs[counter++] = dist2[2*j+1]; + ++n; + } + *start = n; + } + free(dist2); +} - __m128i shuffle = _mm_set_epi64x(0x0101010101010101, 0x0000000000000000); - for (int j = 0; j < QK_K/128; ++j) { +void iq2xs_free_impl(enum ggml_type type) { + GGML_ASSERT(type == GGML_TYPE_IQ2_XXS || type == GGML_TYPE_IQ2_XS || type == GGML_TYPE_IQ1_S || type == GGML_TYPE_IQ1_M || type == GGML_TYPE_IQ2_S); + const int gindex = iq2_data_index(type); + if (iq2_data[gindex].grid) { + free(iq2_data[gindex].grid); iq2_data[gindex].grid = NULL; + free(iq2_data[gindex].map); iq2_data[gindex].map = NULL; + free(iq2_data[gindex].neighbours); iq2_data[gindex].neighbours = NULL; + } +} - const __m128i q4bitsH_0 = _mm_loadu_si128((const __m128i*)qh); qh += 16; - const __m128i q4bitsH_1 = _mm_loadu_si128((const __m128i*)qh); qh += 16; +static int iq2_find_best_neighbour(const uint16_t * restrict neighbours, const uint64_t * restrict grid, + const float * restrict xval, const float * restrict weight, float scale, int8_t * restrict L) { + int num_neighbors = neighbours[0]; + GGML_ASSERT(num_neighbors > 0); + float best_d2 = FLT_MAX; + int grid_index = -1; + for (int j = 1; j <= num_neighbors; ++j) { + const int8_t * pg = (const int8_t *)(grid + neighbours[j]); + float d2 = 0; + for (int i = 0; i < 8; ++i) { + float q = pg[i]; + float diff = scale*q - xval[i]; + d2 += weight[i]*diff*diff; + } + if (d2 < best_d2) { + best_d2 = d2; grid_index = neighbours[j]; + } + } + GGML_ASSERT(grid_index >= 0); + const int8_t * pg = (const int8_t *)(grid + grid_index); + for (int i = 0; i < 8; ++i) L[i] = (pg[i] - 1)/2; + return grid_index; +} - const __m128i q4h_0 = _mm_slli_epi16(_mm_and_si128(q4bitsH_0, m3), 4); - const __m128i q4h_1 = _mm_slli_epi16(_mm_and_si128(q4bitsH_1, m3), 4); - const __m128i q4h_2 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_0, 2), m3), 4); - const __m128i q4h_3 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_1, 2), m3), 4); - const __m128i q4h_4 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_0, 4), m3), 4); - const __m128i q4h_5 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_1, 4), m3), 4); - const __m128i q4h_6 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_0, 6), m3), 4); - const __m128i q4h_7 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH_1, 6), m3), 4); +static void quantize_row_iq2_xxs_impl(const float * restrict x, void * restrict vy, int64_t n, const float * restrict quant_weights) { - const __m128i q4bits1_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4bits1_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4bits2_0 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; - const __m128i q4bits2_1 = _mm_loadu_si128((const __m128i*)q4); q4 += 16; + const int gindex = iq2_data_index(GGML_TYPE_IQ2_XXS); - const __m128i q4_0 = _mm_or_si128(_mm_and_si128(q4bits1_0, m4), q4h_0); - const __m128i q4_1 = _mm_or_si128(_mm_and_si128(q4bits1_1, m4), q4h_1); - const __m128i q4_2 = _mm_or_si128(_mm_and_si128(q4bits2_0, m4), q4h_2); - const __m128i q4_3 = _mm_or_si128(_mm_and_si128(q4bits2_1, m4), q4h_3); - const __m128i q4_4 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_0, 4), m4), q4h_4); - const __m128i q4_5 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits1_1, 4), m4), q4h_5); - const __m128i q4_6 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_0, 4), m4), q4h_6); - const __m128i q4_7 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(q4bits2_1, 4), m4), q4h_7); + const uint64_t * kgrid_q2xs = iq2_data[gindex].grid; + const int * kmap_q2xs = iq2_data[gindex].map; + const uint16_t * kneighbors_q2xs = iq2_data[gindex].neighbours; - const __m128i q8_0 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_1 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_2 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_3 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_4 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_5 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_6 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; - const __m128i q8_7 = _mm_loadu_si128((const __m128i*)q8); q8 += 16; + GGML_ASSERT(quant_weights && "missing quantization weights"); + GGML_ASSERT(kgrid_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kmap_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - __m128i q8s_0 = _mm_maddubs_epi16(m32s, q8_0); - __m128i q8s_1 = _mm_maddubs_epi16(m32s, q8_1); - __m128i q8s_2 = _mm_maddubs_epi16(m32s, q8_2); - __m128i q8s_3 = _mm_maddubs_epi16(m32s, q8_3); - __m128i q8s_4 = _mm_maddubs_epi16(m32s, q8_4); - __m128i q8s_5 = _mm_maddubs_epi16(m32s, q8_5); - __m128i q8s_6 = _mm_maddubs_epi16(m32s, q8_6); - __m128i q8s_7 = _mm_maddubs_epi16(m32s, q8_7); + const int kMaxQ = 3; - __m128i p16_0 = _mm_maddubs_epi16(q4_0, q8_0); - __m128i p16_1 = _mm_maddubs_epi16(q4_1, q8_1); - __m128i p16_2 = _mm_maddubs_epi16(q4_2, q8_2); - __m128i p16_3 = _mm_maddubs_epi16(q4_3, q8_3); - __m128i p16_4 = _mm_maddubs_epi16(q4_4, q8_4); - __m128i p16_5 = _mm_maddubs_epi16(q4_5, q8_5); - __m128i p16_6 = _mm_maddubs_epi16(q4_6, q8_6); - __m128i p16_7 = _mm_maddubs_epi16(q4_7, q8_7); + const int64_t nbl = n/QK_K; - p16_0 = _mm_sub_epi16(p16_0, q8s_0); - p16_1 = _mm_sub_epi16(p16_1, q8s_1); - p16_2 = _mm_sub_epi16(p16_2, q8s_2); - p16_3 = _mm_sub_epi16(p16_3, q8s_3); - p16_4 = _mm_sub_epi16(p16_4, q8s_4); - p16_5 = _mm_sub_epi16(p16_5, q8s_5); - p16_6 = _mm_sub_epi16(p16_6, q8s_6); - p16_7 = _mm_sub_epi16(p16_7, q8s_7); + block_iq2_xxs * y = vy; - const __m128i scale_0 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi8(shuffle, m2); - const __m128i scale_1 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi8(shuffle, m2); - const __m128i scale_2 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi8(shuffle, m2); - const __m128i scale_3 = _mm_shuffle_epi8(scales, shuffle); - shuffle = _mm_add_epi8(shuffle, m2); + float scales[QK_K/32]; + float weight[32]; + float xval[32]; + int8_t L[32]; + int8_t Laux[32]; + float waux[32]; + uint8_t block_signs[4]; + uint32_t q2[2*(QK_K/32)]; - p16_0 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_0), p16_0); - p16_1 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_0, scale_0)), p16_1); - p16_2 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_1), p16_2); - p16_3 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_1, scale_1)), p16_3); - p16_4 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_2), p16_4); - p16_5 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_2, scale_2)), p16_5); - p16_6 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_3), p16_6); - p16_7 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_3, scale_3)), p16_7); + for (int ibl = 0; ibl < nbl; ++ibl) { - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_4, p16_6)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_5, p16_7)); + y[ibl].d = GGML_FP32_TO_FP16(0.f); + memset(q2, 0, QK_K/4); + + float max_scale = 0; + + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = sumx2/QK_K; + + for (int ib = 0; ib < QK_K/32; ++ib) { + const float * xb = xbl + 32*ib; + const float * qw = quant_weights + QK_K*ibl + 32*ib; + for (int i = 0; i < 32; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + for (int i = 0; i < 32; ++i) waux[i] = sqrtf(weight[i]); + for (int k = 0; k < 4; ++k) { + int nflip = 0; + uint8_t s = 0; + for (int i = 0; i < 8; ++i) { + if (xb[8*k + i] >= 0) xval[8*k + i] = xb[8*k + i]; + else { + xval[8*k + i] = -xb[8*k + i]; ++nflip; s |= (1 << i); + } + } + if (nflip%2) { + int imin = 0; float min = weight[8*k+imin]*xb[8*k+imin]*xb[8*k+imin]; + for (int i = 1; i < 8; ++i) { + float ax = weight[8*k+i]*xb[8*k+i]*xb[8*k+i]; + if (ax < min) { + min = ax; imin = i; + } + } + xval[8*k+imin] = -xval[8*k+imin]; + s ^= (1 << imin); + } + block_signs[k] = s & 127; + } + float max = xval[0]; + for (int i = 1; i < 32; ++i) max = MAX(max, xval[i]); + if (!max) { + scales[ib] = 0; + memset(L, 0, 32); + continue; + } + float scale = make_qp_quants(32, kMaxQ+1, xval, (uint8_t*)L, weight); + float eff_max = scale*kMaxQ; + float best = 0; + for (int is = -6; is <= 6; ++is) { + float id = (2*kMaxQ-1+is*0.1f)/eff_max; + float this_scale = 1/id; + for (int k = 0; k < 4; ++k) { + for (int i = 0; i < 8; ++i) { + int l = nearest_int(0.5f*(id*xval[8*k+i]-1)); + Laux[8*k+i] = MAX(0, MIN(kMaxQ-1, l)); + } + uint16_t u = 0; + for (int i = 0; i < 8; ++i) u |= (Laux[8*k+i] << 2*i); + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq2_find_best_neighbour(neighbours, kgrid_q2xs, xval + 8*k, waux + 8*k, this_scale, Laux + 8*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 32; ++i) { + float w = weight[i]; + float q = 2*Laux[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0 && sumqx*sumqx > best*sumq2) { + scale = sumqx/sumq2; best = scale*sumqx; + memcpy(L, Laux, 32); + } + } + if (scale > 0) { + float id = 1/scale; + for (int k = 0; k < 4; ++k) { + uint16_t u = 0; + for (int i = 0; i < 8; ++i) { + int l = nearest_int(0.5f*(id*xval[8*k+i]-1)); + l = MAX(0, MIN(kMaxQ-1, l)); + u |= (l << 2*i); + } + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq2_find_best_neighbour(neighbours, kgrid_q2xs, xval + 8*k, waux + 8*k, scale, L + 8*k); + } + const int8_t * pg = (const int8_t *)(kgrid_q2xs + grid_index); + for (int i = 0; i < 8; ++i) L[8*k+i] = (pg[i] - 1)/2; + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 32; ++i) { + float w = weight[i]; + float q = 2*L[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0) scale = sumqx/sumq2; + } + if (scale < 0) { + // This should never happen, but just in case, flip scale so that it is positive (we use uint's to encode the scale) + // and correspondingly flip quant signs. + scale = -scale; + for (int k = 0; k < 4; ++k) block_signs[k] = (~block_signs[k]) & 127; + } + for (int k = 0; k < 4; ++k) { + uint16_t u = 0; + for (int i = 0; i < 8; ++i) u |= (L[8*k+i] << 2*i); + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + printf("Oops: found point %u not on grid:", u); + for (int i = 0; i < 8; ++i) printf(" %d", L[8*k+i]); + printf("\n"); + GGML_ASSERT(false); + } + q2[2*ib+0] |= (grid_index << 8*k); + q2[2*ib+1] |= (block_signs[k] << 7*k); + } + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + max_scale = MAX(max_scale, scale); + } + if (!max_scale) { + memset(y[ibl].qs, 0, QK_K/4); + continue; } - __m256i sumi = MM256_SET_M128I(sumi_1, sumi_0); - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi)), acc); + float d = max_scale/31; + y[ibl].d = GGML_FP32_TO_FP16(d); + float id = 1/d; + for (int ib = 0; ib < QK_K/32; ++ib) { + int l = nearest_int(0.5f*(id*scales[ib]-1)); + l = MAX(0, MIN(15, l)); + q2[2*ib+1] |= ((uint32_t)l << 28); + } + memcpy(y[ibl].qs, q2, QK_K/4); } +} - *s = hsum_float_8(acc); +static void quantize_row_iq2_xs_impl(const float * restrict x, void * restrict vy, int64_t n, const float * restrict quant_weights) { -#elif defined __riscv_v_intrinsic + const int gindex = iq2_data_index(GGML_TYPE_IQ2_XS); - float sumf = 0; - for (int i = 0; i < nb; ++i) { + const uint64_t * kgrid_q2xs = iq2_data[gindex].grid; + const int * kmap_q2xs = iq2_data[gindex].map; + const uint16_t * kneighbors_q2xs = iq2_data[gindex].neighbours; - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; + GGML_ASSERT(quant_weights && "missing quantization weights"); + GGML_ASSERT(kmap_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kgrid_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; + const int kMaxQ = 3; - const int8_t * restrict scale = x[i].scales; + const int64_t nbl = n/QK_K; - size_t vl; + block_iq2_xs * y = vy; - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); + float scales[QK_K/16]; + float weight[16]; + float xval[16]; + int8_t L[16]; + int8_t Laux[16]; + float waux[16]; + bool is_on_grid[2]; + bool is_on_grid_aux[2]; + uint8_t block_signs[2]; + uint16_t q2[2*(QK_K/16)]; + + for (int ibl = 0; ibl < nbl; ++ibl) { + + y[ibl].d = GGML_FP32_TO_FP16(0.f); + memset(q2, 0, QK_K/4); + memset(y[ibl].scales, 0, QK_K/32); - int sum_t = 0; - int is = 0; + float max_scale = 0; - for (int j = 0; j < QK_K/128; ++j) { + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = sumx2/QK_K; - vl = 32; + for (int ib = 0; ib < QK_K/16; ++ib) { + const float * xb = xbl + 16*ib; + const float * qw = quant_weights + QK_K*ibl + 16*ib; + for (int i = 0; i < 16; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + for (int i = 0; i < 16; ++i) waux[i] = sqrtf(weight[i]); + for (int k = 0; k < 2; ++k) { + int nflip = 0; + uint8_t s = 0; + for (int i = 0; i < 8; ++i) { + if (xb[8*k + i] >= 0) xval[8*k + i] = xb[8*k + i]; + else { + xval[8*k + i] = -xb[8*k + i]; ++nflip; s |= (1 << i); + } + } + if (nflip%2) { + int imin = 0; float min = weight[8*k+imin]*xb[8*k+imin]*xb[8*k+imin]; + for (int i = 1; i < 8; ++i) { + float ax = weight[8*k+i]*xb[8*k+i]*xb[8*k+i]; + if (ax < min) { + min = ax; imin = i; + } + } + xval[8*k+imin] = -xval[8*k+imin]; + s ^= (1 << imin); + } + block_signs[k] = s & 127; + } + float max = xval[0]; + for (int i = 1; i < 16; ++i) max = MAX(max, xval[i]); + if (!max) { + scales[ib] = 0; + memset(L, 0, 16); + continue; + } + float best = 0; + float scale = max/(2*kMaxQ-1); + is_on_grid[0] = is_on_grid[1] = true; + for (int is = -9; is <= 9; ++is) { + float id = (2*kMaxQ-1+is*0.1f)/max; + float this_scale = 1/id; + for (int k = 0; k < 2; ++k) { + for (int i = 0; i < 8; ++i) { + int l = nearest_int(0.5f*(id*xval[8*k+i]-1)); + Laux[8*k+i] = MAX(0, MIN(kMaxQ-1, l)); + } + uint16_t u = 0; + for (int i = 0; i < 8; ++i) u |= (Laux[8*k+i] << 2*i); + int grid_index = kmap_q2xs[u]; + is_on_grid_aux[k] = true; + if (grid_index < 0) { + is_on_grid_aux[k] = false; + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq2_find_best_neighbour(neighbours, kgrid_q2xs, xval + 8*k, waux + 8*k, this_scale, Laux + 8*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 16; ++i) { + float w = weight[i]; + float q = 2*Laux[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0 && sumqx*sumqx > best*sumq2) { + scale = sumqx/sumq2; best = scale*sumqx; + for (int i = 0; i < 16; ++i) L[i] = Laux[i]; + for (int k = 0; k < 2; ++k) is_on_grid[k] = is_on_grid_aux[k]; + } + } + int n_not_ongrid = 0; + for (int k = 0; k < 2; ++k) if (!is_on_grid[k]) ++n_not_ongrid; + if (n_not_ongrid > 0 && scale > 0) { + float id = 1/scale; + for (int k = 0; k < 2; ++k) { + if (is_on_grid[k]) continue; + uint16_t u = 0; + for (int i = 0; i < 8; ++i) { + int l = nearest_int(0.5f*(id*xval[8*k+i]-1)); + l = MAX(0, MIN(kMaxQ-1, l)); + u |= (l << 2*i); + L[8*k + i] = l; + } + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq2_find_best_neighbour(neighbours, kgrid_q2xs, xval + 8*k, waux + 8*k, scale, L + 8*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 16; ++i) { + float w = weight[i]; + float q = 2*L[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0) scale = sumqx/sumq2; + } + if (scale < 0) { + scale = -scale; + for (int k = 0; k < 2; ++k) block_signs[k] = (~block_signs[k]) & 127; + } + for (int k = 0; k < 2; ++k) { + uint16_t u = 0; + for (int i = 0; i < 8; ++i) u |= (L[8*k+i] << 2*i); + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + printf("Oops: found point %u not on grid:", u); + for (int i = 0; i < 8; ++i) printf(" %d", L[8*k+i]); + printf("\n"); + GGML_ASSERT(false); + } + q2[2*ib+k] = grid_index | (block_signs[k] << 9); + } + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + max_scale = MAX(max_scale, scale); + } - // load qh - vuint8m1_t qh_x = __riscv_vle8_v_u8m1(qh, vl); + if (!max_scale) { + memset(y[ibl].qs, 0, QK_K/4); + continue; + } - // load Q6 - vuint8m1_t q6_0 = __riscv_vle8_v_u8m1(q6, vl); - vuint8m1_t q6_1 = __riscv_vle8_v_u8m1(q6+32, vl); + float d = max_scale/31; + y[ibl].d = GGML_FP32_TO_FP16(d); + float id = 1/d; + for (int ib = 0; ib < QK_K/16; ++ib) { + int l = nearest_int(0.5f*(id*scales[ib]-1)); + l = MAX(0, MIN(15, l)); + if (ib%2 == 0) y[ibl].scales[ib/2] = l; + else y[ibl].scales[ib/2] |= (l << 4); + } + memcpy(y[ibl].qs, q2, QK_K/4); - vuint8m1_t q6a_0 = __riscv_vand_vx_u8m1(q6_0, 0x0F, vl); - vuint8m1_t q6a_1 = __riscv_vand_vx_u8m1(q6_1, 0x0F, vl); - vuint8m1_t q6s_0 = __riscv_vsrl_vx_u8m1(q6_0, 0x04, vl); - vuint8m1_t q6s_1 = __riscv_vsrl_vx_u8m1(q6_1, 0x04, vl); + } +} - vuint8m1_t qh_0 = __riscv_vand_vx_u8m1(qh_x, 0x03, vl); - vuint8m1_t qh_1 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x2, vl), 0x03 , vl); - vuint8m1_t qh_2 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x4, vl), 0x03 , vl); - vuint8m1_t qh_3 = __riscv_vand_vx_u8m1(__riscv_vsrl_vx_u8m1(qh_x, 0x6, vl), 0x03 , vl); +size_t quantize_iq2_xxs(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq2_xxs_impl(src, qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += nblock*sizeof(block_iq2_xxs); + } + return nrow * nblock * sizeof(block_iq2_xxs); +} - vuint8m1_t qhi_0 = __riscv_vor_vv_u8m1(q6a_0, __riscv_vsll_vx_u8m1(qh_0, 0x04, vl), vl); - vuint8m1_t qhi_1 = __riscv_vor_vv_u8m1(q6a_1, __riscv_vsll_vx_u8m1(qh_1, 0x04, vl), vl); - vuint8m1_t qhi_2 = __riscv_vor_vv_u8m1(q6s_0, __riscv_vsll_vx_u8m1(qh_2, 0x04, vl), vl); - vuint8m1_t qhi_3 = __riscv_vor_vv_u8m1(q6s_1, __riscv_vsll_vx_u8m1(qh_3, 0x04, vl), vl); +size_t quantize_iq2_xs(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq2_xs_impl(src, qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += nblock*sizeof(block_iq2_xs); + } + return nrow * nblock * sizeof(block_iq2_xs); +} + +// +// ============================================= 3-bit using D4 lattice +// + +typedef struct { + uint32_t * grid; + int * map; + uint16_t * neighbours; +} iq3_entry_t; + +static iq3_entry_t iq3_data[2] = { + {NULL, NULL, NULL}, + {NULL, NULL, NULL}, +}; + +static inline int iq3_data_index(int grid_size) { + (void)grid_size; + GGML_ASSERT(grid_size == 256 || grid_size == 512); + return grid_size == 256 ? 0 : 1; +} + +static int iq3_compare_func(const void * left, const void * right) { + const int * l = (const int *)left; + const int * r = (const int *)right; + return l[0] < r[0] ? -1 : l[0] > r[0] ? 1 : l[1] < r[1] ? -1 : l[1] > r[1] ? 1 : 0; +} - vint8m1_t a_0 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_0), 32, vl); - vint8m1_t a_1 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_1), 32, vl); - vint8m1_t a_2 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_2), 32, vl); - vint8m1_t a_3 = __riscv_vsub_vx_i8m1(__riscv_vreinterpret_v_u8m1_i8m1(qhi_3), 32, vl); +void iq3xs_init_impl(int grid_size) { + const int gindex = iq3_data_index(grid_size); + if (iq3_data[gindex].grid) { + return; + } + static const uint16_t kgrid_256[256] = { + 0, 2, 4, 9, 11, 15, 16, 18, 25, 34, 59, 61, 65, 67, 72, 74, + 81, 85, 88, 90, 97, 108, 120, 128, 130, 132, 137, 144, 146, 153, 155, 159, + 169, 175, 189, 193, 199, 200, 202, 213, 248, 267, 287, 292, 303, 315, 317, 321, + 327, 346, 362, 413, 436, 456, 460, 462, 483, 497, 513, 515, 520, 522, 529, 531, + 536, 538, 540, 551, 552, 576, 578, 585, 592, 594, 641, 643, 648, 650, 657, 664, + 698, 704, 706, 720, 729, 742, 758, 769, 773, 808, 848, 852, 870, 889, 901, 978, + 992, 1024, 1026, 1033, 1035, 1040, 1042, 1046, 1049, 1058, 1089, 1091, 1093, 1096, 1098, 1105, + 1112, 1139, 1143, 1144, 1152, 1154, 1161, 1167, 1168, 1170, 1183, 1184, 1197, 1217, 1224, 1228, + 1272, 1276, 1309, 1323, 1347, 1367, 1377, 1404, 1473, 1475, 1486, 1509, 1537, 1544, 1546, 1553, + 1555, 1576, 1589, 1594, 1600, 1602, 1616, 1625, 1636, 1638, 1665, 1667, 1672, 1685, 1706, 1722, + 1737, 1755, 1816, 1831, 1850, 1856, 1862, 1874, 1901, 1932, 1950, 1971, 2011, 2032, 2052, 2063, + 2077, 2079, 2091, 2095, 2172, 2192, 2207, 2208, 2224, 2230, 2247, 2277, 2308, 2345, 2356, 2389, + 2403, 2424, 2501, 2504, 2506, 2520, 2570, 2593, 2616, 2624, 2630, 2646, 2669, 2700, 2714, 2746, + 2754, 2795, 2824, 2835, 2839, 2874, 2882, 2905, 2984, 3028, 3042, 3092, 3108, 3110, 3124, 3153, + 3185, 3215, 3252, 3288, 3294, 3364, 3397, 3434, 3483, 3523, 3537, 3587, 3589, 3591, 3592, 3610, + 3626, 3670, 3680, 3722, 3749, 3754, 3776, 3789, 3803, 3824, 3857, 3873, 3904, 3906, 3924, 3992, + }; + static const uint16_t kgrid_512[512] = { + 0, 1, 2, 5, 7, 8, 9, 10, 12, 14, 16, 17, 21, 27, 32, 34, + 37, 39, 41, 43, 48, 50, 57, 60, 63, 64, 65, 66, 68, 72, 73, 77, + 80, 83, 87, 89, 93, 100, 113, 117, 122, 128, 129, 133, 135, 136, 139, 142, + 145, 149, 152, 156, 162, 165, 167, 169, 171, 184, 187, 195, 201, 205, 208, 210, + 217, 219, 222, 228, 232, 234, 247, 249, 253, 256, 267, 271, 273, 276, 282, 288, + 291, 297, 312, 322, 324, 336, 338, 342, 347, 353, 357, 359, 374, 379, 390, 393, + 395, 409, 426, 441, 448, 450, 452, 464, 466, 470, 475, 488, 492, 512, 513, 514, + 516, 520, 521, 523, 525, 527, 528, 530, 537, 540, 542, 556, 558, 561, 570, 576, + 577, 579, 582, 584, 588, 593, 600, 603, 609, 616, 618, 632, 638, 640, 650, 653, + 655, 656, 660, 666, 672, 675, 685, 688, 698, 705, 708, 711, 712, 715, 721, 727, + 728, 732, 737, 754, 760, 771, 773, 778, 780, 793, 795, 802, 806, 808, 812, 833, + 840, 843, 849, 856, 858, 873, 912, 916, 919, 932, 934, 961, 963, 968, 970, 977, + 989, 993, 1010, 1016, 1024, 1025, 1027, 1029, 1031, 1032, 1034, 1036, 1038, 1041, 1043, 1047, + 1048, 1050, 1057, 1059, 1061, 1064, 1066, 1079, 1080, 1083, 1085, 1088, 1090, 1096, 1099, 1103, + 1106, 1109, 1113, 1116, 1122, 1129, 1153, 1156, 1159, 1169, 1171, 1176, 1183, 1185, 1195, 1199, + 1209, 1212, 1216, 1218, 1221, 1225, 1234, 1236, 1241, 1243, 1250, 1256, 1270, 1281, 1287, 1296, + 1299, 1306, 1309, 1313, 1338, 1341, 1348, 1353, 1362, 1375, 1376, 1387, 1400, 1408, 1410, 1415, + 1425, 1453, 1457, 1477, 1481, 1494, 1496, 1507, 1512, 1538, 1545, 1547, 1549, 1551, 1554, 1561, + 1563, 1565, 1570, 1572, 1575, 1577, 1587, 1593, 1601, 1603, 1605, 1612, 1617, 1619, 1632, 1648, + 1658, 1662, 1664, 1674, 1680, 1690, 1692, 1704, 1729, 1736, 1740, 1745, 1747, 1751, 1752, 1761, + 1763, 1767, 1773, 1787, 1795, 1801, 1806, 1810, 1817, 1834, 1840, 1844, 1857, 1864, 1866, 1877, + 1882, 1892, 1902, 1915, 1934, 1953, 1985, 1987, 2000, 2002, 2013, 2048, 2052, 2058, 2064, 2068, + 2071, 2074, 2081, 2088, 2104, 2114, 2119, 2121, 2123, 2130, 2136, 2141, 2147, 2153, 2157, 2177, + 2179, 2184, 2189, 2193, 2203, 2208, 2223, 2226, 2232, 2244, 2249, 2251, 2256, 2258, 2265, 2269, + 2304, 2306, 2324, 2335, 2336, 2361, 2373, 2375, 2385, 2418, 2443, 2460, 2480, 2504, 2509, 2520, + 2531, 2537, 2562, 2568, 2572, 2578, 2592, 2596, 2599, 2602, 2614, 2620, 2625, 2627, 2629, 2634, + 2641, 2650, 2682, 2688, 2697, 2707, 2712, 2718, 2731, 2754, 2759, 2760, 2775, 2788, 2793, 2805, + 2811, 2817, 2820, 2832, 2842, 2854, 2890, 2902, 2921, 2923, 2978, 3010, 3012, 3026, 3081, 3083, + 3085, 3097, 3099, 3120, 3136, 3152, 3159, 3188, 3210, 3228, 3234, 3245, 3250, 3256, 3264, 3276, + 3281, 3296, 3349, 3363, 3378, 3392, 3395, 3420, 3440, 3461, 3488, 3529, 3531, 3584, 3588, 3591, + 3600, 3602, 3614, 3616, 3628, 3634, 3650, 3657, 3668, 3683, 3685, 3713, 3716, 3720, 3726, 3729, + 3736, 3753, 3778, 3802, 3805, 3819, 3841, 3845, 3851, 3856, 3880, 3922, 3938, 3970, 3993, 4032, + }; - // load Q8 and take product - vint16m2_t va_q_0 = __riscv_vwmul_vv_i16m2(a_0, __riscv_vle8_v_i8m1(q8, vl), vl); - vint16m2_t va_q_1 = __riscv_vwmul_vv_i16m2(a_1, __riscv_vle8_v_i8m1(q8+32, vl), vl); - vint16m2_t va_q_2 = __riscv_vwmul_vv_i16m2(a_2, __riscv_vle8_v_i8m1(q8+64, vl), vl); - vint16m2_t va_q_3 = __riscv_vwmul_vv_i16m2(a_3, __riscv_vle8_v_i8m1(q8+96, vl), vl); + const int kmap_size = 4096; + const int nwant = grid_size == 256 ? 2 : 3; + const uint16_t * kgrid = grid_size == 256 ? kgrid_256 : kgrid_512; + uint32_t * kgrid_q3xs; + int * kmap_q3xs; + uint16_t * kneighbors_q3xs; + + //printf("================================================================= %s(grid_size = %d)\n", __func__, grid_size); + uint32_t * the_grid = (uint32_t *)malloc(grid_size*sizeof(uint32_t)); + for (int k = 0; k < grid_size; ++k) { + int8_t * pos = (int8_t *)(the_grid + k); + for (int i = 0; i < 4; ++i) { + int l = (kgrid[k] >> 3*i) & 0x7; + pos[i] = 2*l + 1; + } + } + kgrid_q3xs = the_grid; + iq3_data[gindex].grid = the_grid; + kmap_q3xs = (int *)malloc(kmap_size*sizeof(int)); + iq3_data[gindex].map = kmap_q3xs; + for (int i = 0; i < kmap_size; ++i) kmap_q3xs[i] = -1; + uint32_t aux32; + uint8_t * aux8 = (uint8_t *)&aux32; + for (int i = 0; i < grid_size; ++i) { + aux32 = kgrid_q3xs[i]; + uint16_t index = 0; + for (int k=0; k<4; ++k) { + uint16_t q = (aux8[k] - 1)/2; + index |= (q << 3*k); + } + kmap_q3xs[index] = i; + } + int8_t pos[4]; + int * dist2 = (int *)malloc(2*grid_size*sizeof(int)); + int num_neighbors = 0, num_not_in_map = 0; + for (int i = 0; i < kmap_size; ++i) { + if (kmap_q3xs[i] >= 0) continue; + ++num_not_in_map; + for (int k = 0; k < 4; ++k) { + int l = (i >> 3*k) & 0x7; + pos[k] = 2*l + 1; + } + for (int j = 0; j < grid_size; ++j) { + const int8_t * pg = (const int8_t *)(kgrid_q3xs + j); + int d2 = 0; + for (int k = 0; k < 4; ++k) d2 += (pg[k] - pos[k])*(pg[k] - pos[k]); + dist2[2*j+0] = d2; + dist2[2*j+1] = j; + } + qsort(dist2, grid_size, 2*sizeof(int), iq3_compare_func); + int n = 0; int d2 = dist2[0]; + int nhave = 1; + for (int j = 0; j < grid_size; ++j) { + if (dist2[2*j] > d2) { + if (nhave == nwant) break; + d2 = dist2[2*j]; + ++nhave; + } + ++n; + } + num_neighbors += n; + } + //printf("%s: %d neighbours in total\n", __func__, num_neighbors); + kneighbors_q3xs = (uint16_t *)malloc((num_neighbors + num_not_in_map)*sizeof(uint16_t)); + iq3_data[gindex].neighbours = kneighbors_q3xs; + int counter = 0; + for (int i = 0; i < kmap_size; ++i) { + if (kmap_q3xs[i] >= 0) continue; + for (int k = 0; k < 4; ++k) { + int l = (i >> 3*k) & 0x7; + pos[k] = 2*l + 1; + } + for (int j = 0; j < grid_size; ++j) { + const int8_t * pg = (const int8_t *)(kgrid_q3xs + j); + int d2 = 0; + for (int k = 0; k < 4; ++k) d2 += (pg[k] - pos[k])*(pg[k] - pos[k]); + dist2[2*j+0] = d2; + dist2[2*j+1] = j; + } + qsort(dist2, grid_size, 2*sizeof(int), iq3_compare_func); + kmap_q3xs[i] = -(counter + 1); + int d2 = dist2[0]; + uint16_t * start = &kneighbors_q3xs[counter++]; + int n = 0, nhave = 1; + for (int j = 0; j < grid_size; ++j) { + if (dist2[2*j] > d2) { + if (nhave == nwant) break; + d2 = dist2[2*j]; + ++nhave; + } + kneighbors_q3xs[counter++] = dist2[2*j+1]; + ++n; + } + *start = n; + } + free(dist2); +} - vl = 16; +void iq3xs_free_impl(int grid_size) { + GGML_ASSERT(grid_size == 256 || grid_size == 512); + const int gindex = iq3_data_index(grid_size); + if (iq3_data[gindex].grid) { + free(iq3_data[gindex].grid); iq3_data[gindex].grid = NULL; + free(iq3_data[gindex].map); iq3_data[gindex].map = NULL; + free(iq3_data[gindex].neighbours); iq3_data[gindex].neighbours = NULL; + } +} - vint32m2_t vaux_0 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 0), scale[is+0], vl); - vint32m2_t vaux_1 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_0, 1), scale[is+1], vl); - vint32m2_t vaux_2 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 0), scale[is+2], vl); - vint32m2_t vaux_3 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_1, 1), scale[is+3], vl); - vint32m2_t vaux_4 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 0), scale[is+4], vl); - vint32m2_t vaux_5 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_2, 1), scale[is+5], vl); - vint32m2_t vaux_6 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 0), scale[is+6], vl); - vint32m2_t vaux_7 = __riscv_vwmul_vx_i32m2(__riscv_vget_v_i16m2_i16m1(va_q_3, 1), scale[is+7], vl); +static int iq3_find_best_neighbour(const uint16_t * restrict neighbours, const uint32_t * restrict grid, + const float * restrict xval, const float * restrict weight, float scale, int8_t * restrict L) { + int num_neighbors = neighbours[0]; + GGML_ASSERT(num_neighbors > 0); + float best_d2 = FLT_MAX; + int grid_index = -1; + for (int j = 1; j <= num_neighbors; ++j) { + const int8_t * pg = (const int8_t *)(grid + neighbours[j]); + float d2 = 0; + for (int i = 0; i < 4; ++i) { + float q = pg[i]; + float diff = scale*q - xval[i]; + d2 += weight[i]*diff*diff; + } + if (d2 < best_d2) { + best_d2 = d2; grid_index = neighbours[j]; + } + } + GGML_ASSERT(grid_index >= 0); + const int8_t * pg = (const int8_t *)(grid + grid_index); + for (int i = 0; i < 4; ++i) L[i] = (pg[i] - 1)/2; + return grid_index; +} - vint32m1_t isum0 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_0, vaux_1, vl), vzero, vl); - vint32m1_t isum1 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_2, vaux_3, vl), isum0, vl); - vint32m1_t isum2 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_4, vaux_5, vl), isum1, vl); - vint32m1_t isum3 = __riscv_vredsum_vs_i32m2_i32m1(__riscv_vadd_vv_i32m2(vaux_6, vaux_7, vl), isum2, vl); +static void quantize_row_iq3_xxs_impl(int grid_size, const float * restrict x, void * restrict vy, int64_t n, + const float * restrict quant_weights) { - sum_t += __riscv_vmv_x_s_i32m1_i32(isum3); + const int gindex = iq3_data_index(grid_size); - q6 += 64; qh += 32; q8 += 128; is=8; + const uint32_t * kgrid_q3xs = iq3_data[gindex].grid; + const int * kmap_q3xs = iq3_data[gindex].map; + const uint16_t * kneighbors_q3xs = iq3_data[gindex].neighbours; - } + //GGML_ASSERT(quant_weights && "missing quantization weights"); + GGML_ASSERT(kgrid_q3xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kmap_q3xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q3xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - sumf += d * sum_t; + const int kMaxQ = 8; + + const int64_t nbl = n/QK_K; + ggml_fp16_t * dh; + uint8_t * qs; + int block_size; + if (grid_size == 256) { + block_iq3_xxs * y = vy; + dh = &y->d; + qs = y->qs; + block_size = sizeof(block_iq3_xxs); + } else { + block_iq3_s * y = vy; + dh = &y->d; + qs = y->qs; + block_size = sizeof(block_iq3_s); } + int quant_size = block_size - sizeof(ggml_fp16_t); - *s = sumf; + float scales[QK_K/32]; + float weight[32]; + float xval[32]; + int8_t L[32]; + int8_t Laux[32]; + float waux[32]; + bool is_on_grid[8]; + bool is_on_grid_aux[8]; + uint8_t block_signs[8]; + uint8_t q3[3*(QK_K/8)+QK_K/32]; + uint32_t * scales_and_signs = (uint32_t *)(q3 + QK_K/4); + uint8_t * qh = q3 + 3*(QK_K/8); + + for (int ibl = 0; ibl < nbl; ++ibl) { + + dh[0] = GGML_FP32_TO_FP16(0.f); + memset(q3, 0, 3*QK_K/8+QK_K/32); -#else + float max_scale = 0; - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = 2*sumx2/QK_K; - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - for (int j = 0; j < QK_K; j += 128) { - for (int l = 0; l < 32; ++l) { - a[l + 0] = (int8_t)((q4[l + 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32; - a[l + 32] = (int8_t)((q4[l + 32] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32; - a[l + 64] = (int8_t)((q4[l + 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32; - a[l + 96] = (int8_t)((q4[l + 32] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32; + for (int ib = 0; ib < QK_K/32; ++ib) { + const float * xb = xbl + 32*ib; + if (quant_weights) { + const float * qw = quant_weights + QK_K*ibl + 32*ib; + for (int i = 0; i < 32; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + } else { + for (int i = 0; i < 32; ++i) weight[i] = xb[i]*xb[i]; } - a += 128; - q4 += 64; - qh += 32; + for (int i = 0; i < 32; ++i) waux[i] = sqrtf(weight[i]); + for (int k = 0; k < 4; ++k) { + int nflip = 0; + uint8_t s = 0; + for (int i = 0; i < 8; ++i) { + if (xb[8*k + i] >= 0) xval[8*k + i] = xb[8*k + i]; + else { + xval[8*k + i] = -xb[8*k + i]; ++nflip; s |= (1 << i); + } + } + if (nflip%2) { + int imin = 0; float min = weight[8*k+imin]*xb[8*k+imin]*xb[8*k+imin]; + for (int i = 1; i < 8; ++i) { + float ax = weight[8*k+i]*xb[8*k+i]*xb[8*k+i]; + if (ax < min) { + min = ax; imin = i; + } + } + xval[8*k+imin] = -xval[8*k+imin]; + s ^= (1 << imin); + } + block_signs[k] = s & 127; + } + float max = xval[0]; + for (int i = 1; i < 32; ++i) max = MAX(max, xval[i]); + if (!max) { + scales[ib] = 0; + memset(L, 0, 32); + continue; + } + float best = 0; + float scale = max/(2*kMaxQ-1); + for (int is = -15; is <= 15; ++is) { + float id = (2*kMaxQ-1+is*0.2f)/max; + float this_scale = 1/id; + for (int k = 0; k < 8; ++k) { + for (int i = 0; i < 4; ++i) { + int l = nearest_int(0.5f*(id*xval[4*k+i]-1)); + Laux[4*k+i] = MAX(0, MIN(kMaxQ-1, l)); + } + uint16_t u = 0; + for (int i = 0; i < 4; ++i) u |= (Laux[4*k+i] << 3*i); + int grid_index = kmap_q3xs[u]; + is_on_grid_aux[k] = true; + if (grid_index < 0) { + is_on_grid_aux[k] = false; + const uint16_t * neighbours = kneighbors_q3xs - kmap_q3xs[u] - 1; + grid_index = iq3_find_best_neighbour(neighbours, kgrid_q3xs, xval + 4*k, waux + 4*k, this_scale, Laux + 4*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 32; ++i) { + float w = weight[i]; + float q = 2*Laux[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0 && sumqx*sumqx > best*sumq2) { + scale = sumqx/sumq2; best = scale*sumqx; + for (int i = 0; i < 32; ++i) L[i] = Laux[i]; + for (int k = 0; k < 8; ++k) is_on_grid[k] = is_on_grid_aux[k]; + } + } + int n_not_ongrid = 0; + for (int k = 0; k < 8; ++k) if (!is_on_grid[k]) ++n_not_ongrid; + if (n_not_ongrid > 0 && scale > 0) { + float id = 1/scale; + for (int k = 0; k < 8; ++k) { + if (is_on_grid[k]) continue; + uint16_t u = 0; + for (int i = 0; i < 4; ++i) { + int l = nearest_int(0.5f*(id*xval[4*k+i]-1)); + l = MAX(0, MIN(kMaxQ-1, l)); + u |= (l << 3*i); + } + int grid_index = kmap_q3xs[u]; + if (grid_index < 0) { + const uint16_t * neighbours = kneighbors_q3xs - kmap_q3xs[u] - 1; + grid_index = iq3_find_best_neighbour(neighbours, kgrid_q3xs, xval + 4*k, waux + 4*k, scale, L + 4*k); + } + const int8_t * pg = (const int8_t *)(kgrid_q3xs + grid_index); + for (int i = 0; i < 4; ++i) L[4*k+i] = (pg[i] - 1)/2; + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 32; ++i) { + float w = weight[i]; + float q = 2*L[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0) scale = sumqx/sumq2; + } + if (scale < 0) { + // This should never happen, but just in case, flip scale so that it is positive (we use uint's to encode the scale) + // and correspondingly flip quant signs. + scale = -scale; + for (int k = 0; k < 4; ++k) block_signs[k] = (~block_signs[k]) & 127; + } + for (int k = 0; k < 8; ++k) { + uint16_t u = 0; + for (int i = 0; i < 4; ++i) u |= (L[4*k+i] << 3*i); + int grid_index = kmap_q3xs[u]; + if (grid_index < 0) { + printf("Oops: found point %u not on grid:", u); + for (int i = 0; i < 4; ++i) printf(" %d", L[4*k+i]); + printf("\n"); + GGML_ASSERT(false); + } + if (grid_size == 256) { + q3[8*ib+k] = grid_index; + } else { + q3[8*ib+k] = grid_index & 255; + qh[ib] |= ((grid_index >> 8) << k); + } + + } + scales_and_signs[ib] = block_signs[0] | (block_signs[1] << 7) | (block_signs[2] << 14) | (block_signs[3] << 21); + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + max_scale = MAX(max_scale, scale); } - a = aux8; - int is = 0; - for (int j = 0; j < QK_K/16; ++j) { - int scale = x[i].scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; + + if (!max_scale) { + memset(qs, 0, quant_size); + dh += block_size/sizeof(ggml_fp16_t); + qs += block_size; + continue; } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; + + float d = max_scale/31; + dh[0] = GGML_FP32_TO_FP16(d * 1.0125f); // small improvement via this fudge factor + float id = 1/d; + for (int ib = 0; ib < QK_K/32; ++ib) { + int l = nearest_int(0.5f*(id*scales[ib]-1)); + l = MAX(0, MIN(15, l)); + scales_and_signs[ib] |= ((uint32_t)l << 28); + } + memcpy(qs, q3, quant_size); + + dh += block_size/sizeof(ggml_fp16_t); + qs += block_size; + } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif } -#else +size_t quantize_iq3_xxs(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq3_xxs_impl(256, src, qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += nblock*sizeof(block_iq3_xxs); + } + return nrow * nblock * sizeof(block_iq3_xxs); +} -void ggml_vec_dot_q6_K_q8_K(const int n, float * restrict s, const void * restrict vx, const void * restrict vy) { - assert(n % QK_K == 0); +void quantize_row_iq3_xxs(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_iq3_xxs * restrict y = vy; + quantize_row_iq3_xxs_reference(x, y, k); +} - const block_q6_K * restrict x = vx; - const block_q8_K * restrict y = vy; +void quantize_row_iq3_xxs_reference(const float * restrict x, block_iq3_xxs * restrict y, int64_t k) { + assert(k % QK_K == 0); + quantize_row_iq3_xxs_impl(256, x, y, k, NULL); +} - const int nb = n / QK_K; +static void quantize_row_iq3_s_impl(int block_size, const float * restrict x, void * restrict vy, int n, + const float * restrict quant_weights, + float * scales, + float * weight, + float * xval, + int8_t * L, + int8_t * Laux, + float * waux, + bool * is_on_grid, + bool * is_on_grid_aux, + uint8_t * block_signs) { -#ifdef __ARM_NEON + const int gindex = iq3_data_index(512); - float sum = 0; + const uint32_t * kgrid_q3xs = iq3_data[gindex].grid; + const int * kmap_q3xs = iq3_data[gindex].map; + const uint16_t * kneighbors_q3xs = iq3_data[gindex].neighbours; - const uint8x16_t m4b = vdupq_n_u8(0xF); - const int8x16_t m32s = vdupq_n_s8(32); -#if defined(__ARM_FEATURE_DOTPROD) - const int32x4_t vzero = vdupq_n_s32(0); -#endif + //GGML_ASSERT(quant_weights && "missing quantization weights"); + GGML_ASSERT(kgrid_q3xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kmap_q3xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q3xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - const uint8x16_t mone = vdupq_n_u8(3); + const int kMaxQ = 8; - int8x16x4_t q6bytes; - uint8x16x4_t q6h; + const int64_t nbl = n/QK_K; - for (int i = 0; i < nb; ++i) { + block_iq3_s * y = vy; - const float d_all = (float)x[i].d; + const int bs4 = block_size/4; + const int bs8 = block_size/8; - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; + for (int ibl = 0; ibl < nbl; ++ibl) { - const int8_t * restrict scale = x[i].scales; + memset(&y[ibl], 0, sizeof(block_iq3_s)); + y[ibl].d = GGML_FP32_TO_FP16(0.f); - int32_t isum = 0; + uint8_t * qs = y[ibl].qs; + uint8_t * qh = y[ibl].qh; + uint8_t * signs = y[ibl].signs; - uint8x16_t qhbits = vld1q_u8(qh); - uint8x16x2_t q6bits = vld1q_u8_x2(q6); - int8x16x4_t q8bytes = vld1q_s8_x4(q8); + float max_scale = 0; - q6h.val[0] = vshlq_n_u8(vandq_u8(mone, qhbits), 4); - uint8x16_t shifted = vshrq_n_u8(qhbits, 2); - q6h.val[1] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits, 4); - q6h.val[2] = vshlq_n_u8(vandq_u8(mone, shifted), 4); - shifted = vshrq_n_u8(qhbits, 6); - q6h.val[3] = vshlq_n_u8(vandq_u8(mone, shifted), 4); + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = 2*sumx2/QK_K; - q6bytes.val[0] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[0], m4b), q6h.val[0])), m32s); - q6bytes.val[1] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vandq_u8(q6bits.val[1], m4b), q6h.val[1])), m32s); - q6bytes.val[2] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[0], 4), q6h.val[2])), m32s); - q6bytes.val[3] = vsubq_s8(vreinterpretq_s8_u8(vorrq_u8(vshrq_n_u8(q6bits.val[1], 4), q6h.val[3])), m32s); + for (int ib = 0; ib < QK_K/block_size; ++ib) { + const float * xb = xbl + block_size*ib; + if (quant_weights) { + const float * qw = quant_weights + QK_K*ibl + block_size*ib; + for (int i = 0; i < block_size; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + } else { + for (int i = 0; i < block_size; ++i) weight[i] = xb[i]*xb[i]; + } + for (int i = 0; i < block_size; ++i) waux[i] = sqrtf(weight[i]); + for (int k = 0; k < bs8; ++k) { + uint8_t s = 0; + for (int i = 0; i < 8; ++i) { + if (xb[8*k + i] >= 0) xval[8*k + i] = xb[8*k + i]; + else { + xval[8*k + i] = -xb[8*k + i]; s |= (1 << i); + } + } + block_signs[k] = s; + } + float max = xval[0]; + for (int i = 1; i < block_size; ++i) max = MAX(max, xval[i]); + if (!max) { + scales[ib] = 0; + continue; + } + float best = 0; + float scale = max/(2*kMaxQ-1); + for (int k = 0; k < bs4; ++k) is_on_grid[k] = false; + for (int is = -9; is <= 9; ++is) { + float id = (2*kMaxQ-1+is*0.2f)/max; + float this_scale = 1/id; + for (int k = 0; k < bs4; ++k) { + for (int i = 0; i < 4; ++i) { + int l = nearest_int(0.5f*(id*xval[4*k+i]-1)); + Laux[4*k+i] = MAX(0, MIN(kMaxQ-1, l)); + } + uint16_t u = 0; + for (int i = 0; i < 4; ++i) u |= (Laux[4*k+i] << 3*i); + int grid_index = kmap_q3xs[u]; + is_on_grid_aux[k] = true; + if (grid_index < 0) { + is_on_grid_aux[k] = false; + const uint16_t * neighbours = kneighbors_q3xs - kmap_q3xs[u] - 1; + grid_index = iq3_find_best_neighbour(neighbours, kgrid_q3xs, xval + 4*k, waux + 4*k, this_scale, Laux + 4*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < block_size; ++i) { + float w = weight[i]; + float q = 2*Laux[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0 && sumqx*sumqx > best*sumq2) { + scale = sumqx/sumq2; best = scale*sumqx; + for (int i = 0; i < block_size; ++i) L[i] = Laux[i]; + for (int k = 0; k < bs4; ++k) is_on_grid[k] = is_on_grid_aux[k]; + } + } + int n_not_ongrid = 0; + for (int k = 0; k < bs4; ++k) if (!is_on_grid[k]) ++n_not_ongrid; + if (n_not_ongrid > 0 && scale > 0) { + float id = 1/scale; + for (int k = 0; k < bs4; ++k) { + //if (is_on_grid[k]) continue; + uint16_t u = 0; + for (int i = 0; i < 4; ++i) { + int l = nearest_int(0.5f*(id*xval[4*k+i]-1)); + l = MAX(0, MIN(kMaxQ-1, l)); + u |= (l << 3*i); + } + int grid_index = kmap_q3xs[u]; + if (grid_index < 0) { + const uint16_t * neighbours = kneighbors_q3xs - kmap_q3xs[u] - 1; + grid_index = iq3_find_best_neighbour(neighbours, kgrid_q3xs, xval + 4*k, waux + 4*k, scale, L + 4*k); + } + const int8_t * pg = (const int8_t *)(kgrid_q3xs + grid_index); + for (int i = 0; i < 4; ++i) L[4*k+i] = (pg[i] - 1)/2; + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < block_size; ++i) { + float w = weight[i]; + float q = 2*L[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0) scale = sumqx/sumq2; + } + if (scale < 0) { + // This should never happen, but just in case, flip scale so that it is positive (we use uint's to encode the scale) + // and correspondingly flip quant signs. + scale = -scale; + for (int k = 0; k < bs8; ++k) block_signs[k] = ~block_signs[k]; + } + for (int k = 0; k < bs4; ++k) { + uint16_t u = 0; + for (int i = 0; i < 4; ++i) u |= (L[4*k+i] << 3*i); + int grid_index = kmap_q3xs[u]; + if (grid_index < 0) { + printf("Oops: found point %u not on grid:", u); + for (int i = 0; i < 4; ++i) printf(" %d", L[4*k+i]); + printf("\n"); + GGML_ASSERT(false); + } + qs[k] = grid_index & 255; + qh[(ib*bs4+k)/8] |= ((grid_index >> 8) << ((ib*bs4+k)%8)); + } + qs += bs4; + for (int k = 0; k < bs8; ++k) signs[k] = block_signs[k]; + signs += bs8; + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + max_scale = MAX(max_scale, scale); + } -#if defined(__ARM_FEATURE_DOTPROD) + if (!max_scale) { + continue; + } - isum += vaddvq_s32(vdotq_s32(vzero, q6bytes.val[0], q8bytes.val[0])) * scale[0] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[1], q8bytes.val[1])) * scale[1] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[2], q8bytes.val[2])) * scale[2] + - vaddvq_s32(vdotq_s32(vzero, q6bytes.val[3], q8bytes.val[3])) * scale[3]; -#else + float d = max_scale/31; + y[ibl].d = GGML_FP32_TO_FP16(d * 1.033f); + float id = 1/d; + for (int ib = 0; ib < QK_K/block_size; ib += 2) { + int l1 = nearest_int(0.5f*(id*scales[ib+0]-1)); + l1 = MAX(0, MIN(15, l1)); + int l2 = nearest_int(0.5f*(id*scales[ib+1]-1)); + l2 = MAX(0, MIN(15, l2)); + y[ibl].scales[ib/2] = l1 | (l2 << 4); + } - int16x8_t p0 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[0]), vget_low_s8 (q8bytes.val[0])), - vmull_s8(vget_high_s8(q6bytes.val[0]), vget_high_s8(q8bytes.val[0]))); - int16x8_t p1 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[1]), vget_low_s8 (q8bytes.val[1])), - vmull_s8(vget_high_s8(q6bytes.val[1]), vget_high_s8(q8bytes.val[1]))); - isum += vaddvq_s16(p0) * scale[0] + vaddvq_s16(p1) * scale[1]; - - int16x8_t p2 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[2]), vget_low_s8 (q8bytes.val[2])), - vmull_s8(vget_high_s8(q6bytes.val[2]), vget_high_s8(q8bytes.val[2]))); - int16x8_t p3 = vaddq_s16(vmull_s8(vget_low_s8 (q6bytes.val[3]), vget_low_s8 (q8bytes.val[3])), - vmull_s8(vget_high_s8(q6bytes.val[3]), vget_high_s8(q8bytes.val[3]))); - isum += vaddvq_s16(p2) * scale[2] + vaddvq_s16(p3) * scale[3]; -#endif + } +} - sum += isum * d_all * y[i].d; +#define IQ3S_BLOCK_SIZE 32 +size_t quantize_iq3_s(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + int64_t nblock = n_per_row/QK_K; + float scales[QK_K/IQ3S_BLOCK_SIZE]; + float weight[IQ3S_BLOCK_SIZE]; + float xval[IQ3S_BLOCK_SIZE]; + int8_t L[IQ3S_BLOCK_SIZE]; + int8_t Laux[IQ3S_BLOCK_SIZE]; + float waux[IQ3S_BLOCK_SIZE]; + bool is_on_grid[IQ3S_BLOCK_SIZE/4]; + bool is_on_grid_aux[IQ3S_BLOCK_SIZE/4]; + uint8_t block_signs[IQ3S_BLOCK_SIZE/8]; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq3_s_impl(IQ3S_BLOCK_SIZE, src, qrow, n_per_row, quant_weights, + scales, weight, xval, L, Laux, waux, is_on_grid, is_on_grid_aux, block_signs); + src += n_per_row; + qrow += nblock*sizeof(block_iq3_s); + } + return nrow * nblock * sizeof(block_iq3_s); +} - } - *s = sum; +void quantize_row_iq3_s(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_iq3_s * restrict y = vy; + quantize_row_iq3_s_reference(x, y, k); +} -#elif defined __AVX2__ +void quantize_row_iq3_s_reference(const float * restrict x, block_iq3_s * restrict y, int64_t k) { + assert(k % QK_K == 0); + quantize_iq3_s(x, y, 1, k, NULL); +} - const __m256i m4 = _mm256_set1_epi8(0xF); - const __m256i m2 = _mm256_set1_epi8(3); - const __m256i m32s = _mm256_set1_epi8(32); - __m256 acc = _mm256_setzero_ps(); +// =================================== 1.5 bpw =================================================== + +static int iq1_find_best_neighbour(const uint16_t * restrict neighbours, const uint64_t * restrict grid, + const float * restrict xval, const float * restrict weight, float * scale, int8_t * restrict L, int ngrid) { + int num_neighbors = neighbours[0]; + GGML_ASSERT(num_neighbors > 0); + float best_score = 0; + int grid_index = -1; + for (int j = 1; j <= num_neighbors; ++j) { + const int8_t * pg = (const int8_t *)(grid + neighbours[j]); + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 8; ++i) { + float q = (pg[i] - 3)/2; + float w = weight[i]; + sumqx += w*q*xval[i]; + sumq2 += w*q*q; + } + if (sumqx > 0 && sumq2 > 0 && sumqx*sumqx > best_score*sumq2) { + *scale = sumqx/sumq2; best_score = *scale * sumqx; + grid_index = neighbours[j]; + } + } + if (grid_index < 0) { + for (int i = 0; i < ngrid; ++i) { + const int8_t * grid_i = (const int8_t *)(grid + i); + float sumqx = 0, sumq2 = 0; + for (int j = 0; j < 8; ++j) { + float w = weight[j]; + float q = (grid_i[j] - 3)/2; + sumqx += w*q*xval[j]; + sumq2 += w*q*q; + } + if (sumqx > 0 && sumq2 > 0 && sumqx*sumqx > best_score*sumq2) { + *scale = sumqx/sumq2; best_score = *scale*sumqx; + grid_index = i; + } + } + } + if (grid_index < 0) { + printf("Oops, did not find grid point\n"); + printf("Have %d neighbours\n", num_neighbors); + for (int j = 1; j <= num_neighbors; ++j) { + const int8_t * pg = (const int8_t *)(grid + neighbours[j]); + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 8; ++i) { + float q = (pg[i] - 3)/2; + float w = weight[i]; + sumqx += w*q*xval[i]; + sumq2 += w*q*q; + } + printf(" neighbour %d: sumqx = %g sumq2 = %g\n", j, (double)sumqx, (double)sumq2); + } + } + GGML_ASSERT(grid_index >= 0); + //!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! + *scale *= 1.05f; // This is a fudge factor. Don't ask me why it improves the result. + //!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! + const int8_t * pg = (const int8_t *)(grid + grid_index); + for (int i = 0; i < 8; ++i) L[i] = (pg[i] - 1)/2; + return grid_index; +} - for (int i = 0; i < nb; ++i) { +static int iq1_find_best_neighbour2(const uint16_t * restrict neighbours, const uint64_t * restrict grid, + const float * restrict xval, const float * restrict weight, float scale, const float * restrict xg, int8_t * restrict L, int ngrid) { + int num_neighbors = neighbours[0]; + GGML_ASSERT(num_neighbors > 0); + float best_score = FLT_MAX; + int grid_index = -1; + for (int j = 1; j <= num_neighbors; ++j) { + const int8_t * pg = (const int8_t *)(grid + neighbours[j]); + float d2 = 0; + for (int i = 0; i < 8; ++i) { + float q = xg[(pg[i] - 1)/2]; + float w = weight[i]; + float diff = scale*q - xval[i]; + d2 += w*diff*diff; + } + if (d2 < best_score) { + best_score = d2; + grid_index = neighbours[j]; + } + } + if (grid_index < 0) { + for (int i = 0; i < ngrid; ++i) { + const int8_t * grid_i = (const int8_t *)(grid + i); + float d2 = 0; + for (int j = 0; j < 8; ++j) { + float w = weight[j]; + float q = xg[(grid_i[j] - 1)/2]; + float diff = scale*q - xval[i]; + d2 += w*diff*diff; + } + if (d2 < best_score) { + best_score = d2; + grid_index = i; + } + } + } + if (grid_index < 0) { + printf("Oops, did not find grid point\n"); + printf("Have %d neighbours\n", num_neighbors); + for (int j = 1; j <= num_neighbors; ++j) { + const int8_t * pg = (const int8_t *)(grid + neighbours[j]); + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 8; ++i) { + float q = xg[(pg[i] - 1)/2]; + float w = weight[i]; + sumqx += w*q*xval[i]; + sumq2 += w*q*q; + } + printf(" neighbour %d: sumqx = %g sumq2 = %g\n", j, (double)sumqx, (double)sumq2); + } + } + GGML_ASSERT(grid_index >= 0); + const int8_t * pg = (const int8_t *)(grid + grid_index); + for (int i = 0; i < 8; ++i) L[i] = (pg[i] - 1)/2; + return grid_index; +} - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); +static int iq1_sort_helper(const void * left, const void * right) { + const float * l = left; + const float * r = right; + return *l < *r ? -1 : *l > *r ? 1 : 0; +} - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; +#define IQ1S_BLOCK_SIZE 32 +#define IQ1M_BLOCK_SIZE 16 +static void quantize_row_iq1_s_impl(const float * restrict x, void * restrict vy, int64_t n, const float * restrict quant_weights, + float * scales, + float * weight, + float * sumx, + float * sumw, + float * pairs, + int8_t * L, + uint16_t * index, + int8_t * shifts) { - const __m64 scales_1 = _mm_set1_pi8(x[i].scales[0]); - const __m64 scales_2 = _mm_set1_pi8(x[i].scales[1]); - const __m64 scales_3 = _mm_set1_pi8(x[i].scales[2]); - const __m64 scales_4 = _mm_set1_pi8(x[i].scales[3]); + const int gindex = iq2_data_index(GGML_TYPE_IQ1_S); - __m256i sumi = _mm256_setzero_si256(); + const uint64_t * kgrid_q2xs = iq2_data[gindex].grid; + const int * kmap_q2xs = iq2_data[gindex].map; + const uint16_t * kneighbors_q2xs = iq2_data[gindex].neighbours; - const __m128i scale_0 = _mm_set_epi64(scales_2, scales_1); - const __m128i scale_1 = _mm_set_epi64(scales_4, scales_3); + GGML_ASSERT(quant_weights && "missing quantization weights"); + GGML_ASSERT(kgrid_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kmap_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); - const __m128i q4bitsH = _mm_loadu_si128((const __m128i*)qh); + block_iq1_s * y = vy; - const __m256i q4h_0 = _mm256_slli_epi16(_mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q4bitsH, 2), q4bitsH), m2), 4); - const __m256i q4h_1 = _mm256_slli_epi16(_mm256_and_si256(MM256_SET_M128I(_mm_srli_epi16(q4bitsH, 6), _mm_srli_epi16(q4bitsH, 4)), m2), 4); + const int64_t nbl = n/QK_K; - const __m256i q4_0 = _mm256_or_si256(_mm256_and_si256(q4bits1, m4), q4h_0); - const __m256i q4_1 = _mm256_or_si256(_mm256_and_si256(_mm256_srli_epi16(q4bits1, 4), m4), q4h_1); + const int block_size = IQ1S_BLOCK_SIZE; - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + const float x_p[3] = {-1 + IQ1S_DELTA, IQ1S_DELTA, 1 + IQ1S_DELTA}; + const float x_m[3] = {-1 - IQ1S_DELTA, -IQ1S_DELTA, 1 - IQ1S_DELTA}; - __m256i q8s_0 = _mm256_maddubs_epi16(m32s, q8_0); - __m256i q8s_1 = _mm256_maddubs_epi16(m32s, q8_1); - __m256i p16_0 = _mm256_maddubs_epi16(q4_0, q8_0); - __m256i p16_1 = _mm256_maddubs_epi16(q4_1, q8_1); + int * idx = (int *)(pairs + 1); - p16_0 = _mm256_sub_epi16(p16_0, q8s_0); - p16_1 = _mm256_sub_epi16(p16_1, q8s_1); + for (int ibl = 0; ibl < nbl; ++ibl) { - p16_0 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_0), p16_0); - p16_1 = _mm256_madd_epi16(_mm256_cvtepi8_epi16(scale_1), p16_1); + y[ibl].d = GGML_FP32_TO_FP16(0.f); + memset(y[ibl].qs, 0, QK_K/8); + memset(y[ibl].qh, 0, QK_K/16); + + float max_scale = 0; + + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = 2*sumx2/QK_K; + + for (int ib = 0; ib < QK_K/block_size; ++ib) { + const float * xb = xbl + block_size*ib; + const float * qw = quant_weights + QK_K*ibl + block_size*ib; + for (int i = 0; i < block_size; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + float max = fabsf(xb[0]); + for (int i = 1; i < block_size; ++i) max = MAX(max, fabsf(xb[i])); + if (!max) { + scales[ib] = 0; + memset(L, 1, block_size); + continue; + } + // Here we solve exactly the sum of squared difference (SSD) weighted minimization problem. + // With just 3 allowed quant values (-1, 0, 1), we can search exhaustively for the two + // boundaries that split the weights xb[i] into 3 groups. To do so, we sort the weights + // in ascending order, compute Si = sum[weight[j] xb[j], j = 0...i] and + // Wi = sum[weight[j], j = 0...i], and use these to quckly get get the optimum scale + // for each possible and score for each split. + for (int j = 0; j < block_size; ++j) { + pairs[2*j] = xb[j]; + idx[2*j] = j; + } + qsort(pairs, block_size, 2*sizeof(float), iq1_sort_helper); + { + sumx[0] = sumw[0] = 0; + for (int j = 0; j < block_size; ++j) { + int i = idx[2*j]; + sumx[j+1] = sumx[j] + weight[i]*xb[i]; + sumw[j+1] = sumw[j] + weight[i]; + } + } + float best_score = 0, scale = max; + int besti1 = -1, besti2 = -1, best_shift = 0; + for (int i1 = 0; i1 <= block_size; ++i1) { + for (int i2 = i1; i2 <= block_size; ++i2) { + float sumqx = (sumx[i1] - sumx[0])*x_p[0] + (sumx[i2] - sumx[i1])*x_p[1] + (sumx[block_size] - sumx[i2])*x_p[2]; + float sumq2 = (sumw[i1] - sumw[0])*x_p[0]*x_p[0] + (sumw[i2] - sumw[i1])*x_p[1]*x_p[1] + (sumw[block_size] - sumw[i2])*x_p[2]*x_p[2]; + if (sumq2 > 0 && sumqx*sumqx > best_score*sumq2) { + scale = sumqx/sumq2; best_score = scale*sumqx; + besti1 = i1; besti2 = i2; best_shift = 1; + } + sumqx = (sumx[i1] - sumx[0])*x_m[0] + (sumx[i2] - sumx[i1])*x_m[1] + (sumx[block_size] - sumx[i2])*x_m[2]; + sumq2 = (sumw[i1] - sumw[0])*x_m[0]*x_m[0] + (sumw[i2] - sumw[i1])*x_m[1]*x_m[1] + (sumw[block_size] - sumw[i2])*x_m[2]*x_m[2]; + if (sumq2 > 0 && sumqx*sumqx > best_score*sumq2) { + scale = sumqx/sumq2; best_score = scale*sumqx; + besti1 = i1; besti2 = i2; best_shift = -1; + } + } + } + GGML_ASSERT(besti1 >= 0 && besti2 >= 0 && best_shift != 0); + for (int j = 0; j < besti1; ++j) L[idx[2*j]] = 0; + for (int j = besti1; j < besti2; ++j) L[idx[2*j]] = 1; + for (int j = besti2; j < block_size; ++j) L[idx[2*j]] = 2; + if (scale < 0) { + for (int j = 0; j < block_size; ++j) L[j] = 2 - L[j]; + scale = -scale; best_shift = -best_shift; + } + bool all_on_grid = true; + const float * xx = best_shift == 1 ? x_p : x_m; + for (int k = 0; k < block_size/8; ++k) { + uint16_t u = 0; + for (int j = 0; j < 8; ++j) u |= (L[8*k+j] << 2*j); + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + all_on_grid = false; + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq1_find_best_neighbour2(neighbours, kgrid_q2xs, xb + 8*k, weight + 8*k, scale, xx, L + 8*k, NGRID_IQ1S); + GGML_ASSERT(grid_index >= 0); + } + index[k] = grid_index; + } + if (!all_on_grid) { + float sumqx = 0, sumq2 = 0; + for (int k = 0; k < block_size/8; ++k) { + const int8_t * pg = (const int8_t *)(kgrid_q2xs + index[k]); + for (int j = 0; j < 8; ++j) { + float w = weight[8*k + j]; + float q = xx[(pg[j] - 1)/2]; + sumqx += w*q*xb[8*k+j]; + sumq2 += w*q*q; + } + } + if (sumqx > 0 && sumq2 > 0) scale = sumqx/sumq2; + } + uint16_t h = 0; + for (int k = 0; k < block_size/8; ++k) { + y[ibl].qs[(block_size/8)*ib + k] = index[k] & 255; + h |= (index[k] >> 8) << 3*k; + } + y[ibl].qh[ib] = h; + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + shifts[ib] = best_shift; + max_scale = MAX(max_scale, scale); + } - sumi = _mm256_add_epi32(sumi, _mm256_add_epi32(p16_0, p16_1)); + if (!max_scale) { + continue; + } - acc = _mm256_fmadd_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(sumi), acc); + float d = max_scale/15; + y[ibl].d = GGML_FP32_TO_FP16(d*1.125f); // 1.125f is another fudge factor. Don't ask me why it is needed. + float id = 1/d; + for (int ib = 0; ib < QK_K/block_size; ++ib) { + int l = nearest_int(0.5f*(id*scales[ib]-1)); + l = MAX(0, MIN(7, l)); + if (shifts[ib] == -1) l |= 8; + y[ibl].qh[ib] |= (l << 12); + } } +} - *s = hsum_float_8(acc); - -#elif defined __AVX__ +size_t quantize_iq1_s(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + float scales[QK_K/IQ1S_BLOCK_SIZE]; + float weight[IQ1S_BLOCK_SIZE]; + int8_t L[IQ1S_BLOCK_SIZE]; + float sumx[IQ1S_BLOCK_SIZE+1]; + float sumw[IQ1S_BLOCK_SIZE+1]; + float pairs[2*IQ1S_BLOCK_SIZE]; + uint16_t index[IQ1S_BLOCK_SIZE/8]; + int8_t shifts[QK_K/IQ1S_BLOCK_SIZE]; + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq1_s_impl(src, qrow, n_per_row, quant_weights, scales, weight, sumx, sumw, pairs, L, index, shifts); + src += n_per_row; + qrow += nblock*sizeof(block_iq1_s); + } + return nrow * nblock * sizeof(block_iq1_s); +} - const __m128i m4 = _mm_set1_epi8(0xF); - const __m128i m2 = _mm_set1_epi8(3); - const __m128i m32s = _mm_set1_epi8(32); +static void quantize_row_iq1_m_impl(const float * restrict x, void * restrict vy, int64_t n, const float * restrict quant_weights, + float * scales, + float * weight, + float * pairs, + int8_t * L, + uint16_t * index, + int8_t * shifts) { - __m256 acc = _mm256_setzero_ps(); + const int gindex = iq2_data_index(GGML_TYPE_IQ1_M); - for (int i = 0; i < nb; ++i) { + const uint64_t * kgrid_q2xs = iq2_data[gindex].grid; + const int * kmap_q2xs = iq2_data[gindex].map; + const uint16_t * kneighbors_q2xs = iq2_data[gindex].neighbours; - const float d = y[i].d * GGML_FP16_TO_FP32(x[i].d); + //GGML_ASSERT(quant_weights && "missing quantization weights"); + GGML_ASSERT(kgrid_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kmap_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; + block_iq1_m * y = vy; - const __m64 scales_1 = _mm_set1_pi8(x[i].scales[0]); - const __m64 scales_2 = _mm_set1_pi8(x[i].scales[1]); - const __m64 scales_3 = _mm_set1_pi8(x[i].scales[2]); - const __m64 scales_4 = _mm_set1_pi8(x[i].scales[3]); + const int64_t nbl = n/QK_K; - __m128i sumi_0 = _mm_setzero_si128(); - __m128i sumi_1 = _mm_setzero_si128(); + const int block_size = IQ1M_BLOCK_SIZE; - const __m128i scale_0 = _mm_set_epi64(scales_2, scales_1); - const __m128i scale_1 = _mm_set_epi64(scales_4, scales_3); + const float x_p[3] = {-1 + IQ1M_DELTA, IQ1M_DELTA, 1 + IQ1M_DELTA}; + const float x_m[3] = {-1 - IQ1M_DELTA, -IQ1M_DELTA, 1 - IQ1M_DELTA}; + const uint8_t masks[4] = {0x00, 0x80, 0x08, 0x88}; - const __m256i q4bits1 = _mm256_loadu_si256((const __m256i*)q4); - const __m128i q4bitsH = _mm_loadu_si128((const __m128i*)qh); + int * idx = (int *)(pairs + 1); - const __m128i q4h_0 = _mm_slli_epi16(_mm_and_si128(q4bitsH, m2), 4); - const __m128i q4h_1 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH, 2), m2), 4); - const __m128i q4h_2 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH, 4), m2), 4); - const __m128i q4h_3 = _mm_slli_epi16(_mm_and_si128(_mm_srli_epi16(q4bitsH, 6), m2), 4); + float sumqx[4], sumq2[4]; - const __m128i q4_0 = _mm_or_si128(_mm_and_si128(_mm256_extractf128_si256(q4bits1, 0), m4), q4h_0); - const __m128i q4_1 = _mm_or_si128(_mm_and_si128(_mm256_extractf128_si256(q4bits1, 1), m4), q4h_1); - const __m128i q4_2 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q4bits1, 0), 4), m4), q4h_2); - const __m128i q4_3 = _mm_or_si128(_mm_and_si128(_mm_srli_epi16(_mm256_extractf128_si256(q4bits1, 1), 4), m4), q4h_3); + iq1m_scale_t s; + const float * xx; - const __m256i q8_0 = _mm256_loadu_si256((const __m256i*)(q8+ 0)); - const __m256i q8_1 = _mm256_loadu_si256((const __m256i*)(q8+32)); + for (int ibl = 0; ibl < nbl; ++ibl) { - __m128i q8s_0 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_0, 0)); - __m128i q8s_1 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_0, 1)); - __m128i q8s_2 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_1, 0)); - __m128i q8s_3 = _mm_maddubs_epi16(m32s, _mm256_extractf128_si256(q8_1, 1)); +#if QK_K == 64 + y[ibl].d = GGML_FP32_TO_FP16(0.f); +#endif + memset(y[ibl].qs, 0, QK_K/8); + memset(y[ibl].qh, 0, QK_K/16); + memset(y[ibl].scales, 0, QK_K/32); - __m128i p16_0 = _mm_maddubs_epi16(q4_0, _mm256_extractf128_si256(q8_0, 0)); - __m128i p16_1 = _mm_maddubs_epi16(q4_1, _mm256_extractf128_si256(q8_0, 1)); - __m128i p16_2 = _mm_maddubs_epi16(q4_2, _mm256_extractf128_si256(q8_1, 0)); - __m128i p16_3 = _mm_maddubs_epi16(q4_3, _mm256_extractf128_si256(q8_1, 1)); + float max_scale = 0; - p16_0 = _mm_sub_epi16(p16_0, q8s_0); - p16_1 = _mm_sub_epi16(p16_1, q8s_1); - p16_2 = _mm_sub_epi16(p16_2, q8s_2); - p16_3 = _mm_sub_epi16(p16_3, q8s_3); + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = 2*sumx2/QK_K; - p16_0 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_0), p16_0); - p16_1 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_0, scale_0)), p16_1); - p16_2 = _mm_madd_epi16(_mm_cvtepi8_epi16(scale_1), p16_2); - p16_3 = _mm_madd_epi16(_mm_cvtepi8_epi16(_mm_unpackhi_epi64(scale_1, scale_1)), p16_3); + for (int ib = 0; ib < QK_K/block_size; ++ib) { + const float * xb = xbl + block_size*ib; + if (quant_weights) { + const float * qw = quant_weights + QK_K*ibl + block_size*ib; + for (int i = 0; i < block_size; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + } else { + for (int i = 0; i < block_size; ++i) weight[i] = xb[i]*xb[i]; + } + float max = fabsf(xb[0]); + for (int i = 1; i < block_size; ++i) max = MAX(max, fabsf(xb[i])); + if (!max) { + scales[ib] = 0; + memset(L, 1, block_size); + continue; + } + // Here we solve exactly the sum of squared difference (SSD) weighted minimization problem. + // With just 3 allowed quant values (-1, 0, 1), we can search exhaustively for the two + // boundaries that split the weights xb[i] into 3 groups. To do so, we sort the weights + // in ascending order, compute Si = sum[weight[j] xb[j], j = 0...i] and + // Wi = sum[weight[j], j = 0...i], and use these to quckly get get the optimum scale + // for each possible and score for each split. + for (int j = 0; j < block_size; ++j) { + pairs[2*j] = xb[j]; + idx[2*j] = j; + } + qsort(pairs, block_size, 2*sizeof(float), iq1_sort_helper); + float best_score = 0, scale = max; + int besti1 = -1, besti2 = -1, best_k = -1; + // 0: +, + + // 1: +, - + // 2: -, + + // 3: -, - + for (int i1 = 0; i1 <= block_size; ++i1) { + for (int i2 = i1; i2 <= block_size; ++i2) { + memset(sumqx, 0, 4*sizeof(float)); + memset(sumq2, 0, 4*sizeof(float)); + for (int j = 0; j < i1; ++j) { + int i = idx[2*j]; + if (i < block_size/2) { + sumqx[0] += weight[i]*x_p[0]*xb[i]; + sumqx[1] += weight[i]*x_p[0]*xb[i]; + sumqx[2] += weight[i]*x_m[0]*xb[i]; + sumqx[3] += weight[i]*x_m[0]*xb[i]; + sumq2[0] += weight[i]*x_p[0]*x_p[0]; + sumq2[1] += weight[i]*x_p[0]*x_p[0]; + sumq2[2] += weight[i]*x_m[0]*x_m[0]; + sumq2[3] += weight[i]*x_m[0]*x_m[0]; + } else { + sumqx[0] += weight[i]*x_p[0]*xb[i]; + sumqx[2] += weight[i]*x_p[0]*xb[i]; + sumqx[1] += weight[i]*x_m[0]*xb[i]; + sumqx[3] += weight[i]*x_m[0]*xb[i]; + sumq2[0] += weight[i]*x_p[0]*x_p[0]; + sumq2[2] += weight[i]*x_p[0]*x_p[0]; + sumq2[1] += weight[i]*x_m[0]*x_m[0]; + sumq2[3] += weight[i]*x_m[0]*x_m[0]; + } + } + for (int j = i1; j < i2; ++j) { + int i = idx[2*j]; + if (i < block_size/2) { + sumqx[0] += weight[i]*x_p[1]*xb[i]; + sumqx[1] += weight[i]*x_p[1]*xb[i]; + sumqx[2] += weight[i]*x_m[1]*xb[i]; + sumqx[3] += weight[i]*x_m[1]*xb[i]; + sumq2[0] += weight[i]*x_p[1]*x_p[1]; + sumq2[1] += weight[i]*x_p[1]*x_p[1]; + sumq2[2] += weight[i]*x_m[1]*x_m[1]; + sumq2[3] += weight[i]*x_m[1]*x_m[1]; + } else { + sumqx[0] += weight[i]*x_p[1]*xb[i]; + sumqx[2] += weight[i]*x_p[1]*xb[i]; + sumqx[1] += weight[i]*x_m[1]*xb[i]; + sumqx[3] += weight[i]*x_m[1]*xb[i]; + sumq2[0] += weight[i]*x_p[1]*x_p[1]; + sumq2[2] += weight[i]*x_p[1]*x_p[1]; + sumq2[1] += weight[i]*x_m[1]*x_m[1]; + sumq2[3] += weight[i]*x_m[1]*x_m[1]; + } + } + for (int j = i2; j < block_size; ++j) { + int i = idx[2*j]; + if (i < block_size/2) { + sumqx[0] += weight[i]*x_p[2]*xb[i]; + sumqx[1] += weight[i]*x_p[2]*xb[i]; + sumqx[2] += weight[i]*x_m[2]*xb[i]; + sumqx[3] += weight[i]*x_m[2]*xb[i]; + sumq2[0] += weight[i]*x_p[2]*x_p[2]; + sumq2[1] += weight[i]*x_p[2]*x_p[2]; + sumq2[2] += weight[i]*x_m[2]*x_m[2]; + sumq2[3] += weight[i]*x_m[2]*x_m[2]; + } else { + sumqx[0] += weight[i]*x_p[2]*xb[i]; + sumqx[2] += weight[i]*x_p[2]*xb[i]; + sumqx[1] += weight[i]*x_m[2]*xb[i]; + sumqx[3] += weight[i]*x_m[2]*xb[i]; + sumq2[0] += weight[i]*x_p[2]*x_p[2]; + sumq2[2] += weight[i]*x_p[2]*x_p[2]; + sumq2[1] += weight[i]*x_m[2]*x_m[2]; + sumq2[3] += weight[i]*x_m[2]*x_m[2]; + } + } + for (int k = 0; k < 4; ++k) { + if (sumq2[k] > 0 && sumqx[k]*sumqx[k] > best_score*sumq2[k]) { + scale = sumqx[k]/sumq2[k]; best_score = scale*sumqx[k]; + besti1 = i1; besti2 = i2; best_k = k; + } + } + } + } + GGML_ASSERT(besti1 >= 0 && besti2 >= 0 && best_k >= 0); + for (int j = 0; j < besti1; ++j) L[idx[2*j]] = 0; + for (int j = besti1; j < besti2; ++j) L[idx[2*j]] = 1; + for (int j = besti2; j < block_size; ++j) L[idx[2*j]] = 2; + if (scale < 0) { + for (int j = 0; j < block_size; ++j) L[j] = 2 - L[j]; + scale = -scale; + best_k = best_k == 0 ? 3 : best_k == 1 ? 2 : best_k == 2 ? 1 : 0; + } + bool all_on_grid = true; + for (int k = 0; k < block_size/8; ++k) { + if (k == 0) xx = best_k < 2 ? x_p : x_m; + else xx = best_k%2 == 0 ? x_p : x_m; + uint16_t u = 0; + for (int j = 0; j < 8; ++j) u |= (L[8*k+j] << 2*j); + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + all_on_grid = false; + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq1_find_best_neighbour2(neighbours, kgrid_q2xs, xb + 8*k, weight + 8*k, scale, xx, L + 8*k, NGRID_IQ1S); + GGML_ASSERT(grid_index >= 0); + } + index[k] = grid_index; + } + if (!all_on_grid) { + float sumqx_f = 0, sumq2_f = 0; + for (int k = 0; k < block_size/8; ++k) { + if (k == 0) xx = best_k < 2 ? x_p : x_m; + else xx = best_k%2 == 0 ? x_p : x_m; + const int8_t * pg = (const int8_t *)(kgrid_q2xs + index[k]); + for (int j = 0; j < 8; ++j) { + float w = weight[8*k + j]; + float q = xx[(pg[j] - 1)/2]; + sumqx_f += w*q*xb[8*k+j]; + sumq2_f += w*q*q; + } + } + if (sumqx_f > 0 && sumq2_f > 0) scale = sumqx_f/sumq2_f; + } + y[ibl].qs[2*ib + 0] = index[0] & 255; + y[ibl].qs[2*ib + 1] = index[1] & 255; + y[ibl].qh[ib] = (index[0] >> 8) | ((index[1] >> 8) << 4); + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + shifts[ib] = best_k; + max_scale = MAX(max_scale, scale); + } - sumi_0 = _mm_add_epi32(sumi_0, _mm_add_epi32(p16_0, p16_2)); - sumi_1 = _mm_add_epi32(sumi_1, _mm_add_epi32(p16_1, p16_3)); + if (!max_scale) { + continue; + } - acc = _mm256_add_ps(_mm256_mul_ps(_mm256_broadcast_ss(&d), _mm256_cvtepi32_ps(MM256_SET_M128I(sumi_1, sumi_0))), acc); + uint16_t * sc = (uint16_t *)y[ibl].scales; +#if QK_K == 64 + float d = max_scale/31; +#else + float d = max_scale/15; +#endif + float id = 1/d; + float sumqx_f = 0, sumq2_f = 0; + for (int ib = 0; ib < QK_K/block_size; ++ib) { + int l = nearest_int(0.5f*(id*scales[ib+0]-1)); +#if QK_K == 64 + l = MAX(0, MIN(15, l)); + sc[ib/4] |= (l << 4*(ib%4)); +#else + l = MAX(0, MIN(7, l)); + sc[ib/4] |= (l << 3*(ib%4)); +#endif + y[ibl].qh[ib] |= masks[shifts[ib]]; + const float * xb = xbl + block_size*ib; + if (quant_weights) { + const float * qw = quant_weights + QK_K*ibl + block_size*ib; + for (int i = 0; i < block_size; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + } else { + for (int i = 0; i < block_size; ++i) weight[i] = xb[i]*xb[i]; + } + for (int k = 0; k < block_size/8; ++k) { + if (k == 0) xx = shifts[ib] < 2 ? x_p : x_m; + else xx = shifts[ib]%2 == 0 ? x_p : x_m; + const int8_t * pg = (const int8_t *)(kgrid_q2xs + y[ibl].qs[2*ib+k] + ((y[ibl].qh[ib] << (8 - 4*k)) & 0x700)); + for (int j = 0; j < 8; ++j) { + float w = weight[8*k + j]; + float q = xx[(pg[j] - 1)/2]*(2*l+1); + sumqx_f += w*q*xb[8*k+j]; + sumq2_f += w*q*q; + } + } + } + if (sumq2_f > 0) d = sumqx_f/sumq2_f; + s.f16 = GGML_FP32_TO_FP16(d*1.1125f); // 1.1125f is another fudge factor. Don't ask me why it is needed. +#if QK_K == 64 + y[ibl].d = s.f16; +#else + sc[0] |= ((s.u16 & 0x000f) << 12); + sc[1] |= ((s.u16 & 0x00f0) << 8); + sc[2] |= ((s.u16 & 0x0f00) << 4); + sc[3] |= ((s.u16 & 0xf000) << 0); +#endif } +} - *s = hsum_float_8(acc); +size_t quantize_iq1_m(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + float scales[QK_K/IQ1M_BLOCK_SIZE]; + float weight[IQ1M_BLOCK_SIZE]; + int8_t L[IQ1M_BLOCK_SIZE]; + float pairs[2*IQ1M_BLOCK_SIZE]; + uint16_t index[IQ1M_BLOCK_SIZE/8]; + int8_t shifts[QK_K/IQ1M_BLOCK_SIZE]; + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq1_m_impl(src, qrow, n_per_row, quant_weights, scales, weight, pairs, L, index, shifts); + src += n_per_row; + qrow += nblock*sizeof(block_iq1_m); + } + return nrow * nblock * sizeof(block_iq1_m); +} -#elif defined __riscv_v_intrinsic +// ============================ 4-bit non-linear quants - float sumf = 0; +static inline int best_index_int8(int n, const int8_t * val, float x) { + if (x <= val[0]) return 0; + if (x >= val[n-1]) return n-1; + int ml = 0, mu = n-1; + while (mu-ml > 1) { + int mav = (ml+mu)/2; + if (x < val[mav]) mu = mav; else ml = mav; + } + return x - val[mu-1] < val[mu] - x ? mu-1 : mu; +} - for (int i = 0; i < nb; ++i) { +static void quantize_row_iq4_nl_impl(const int super_block_size, const int block_size, const float * restrict x, + ggml_fp16_t * dh, uint8_t * q4, uint16_t * scales_h, uint8_t * scales_l, + float * scales, float * weight, uint8_t * L, + const int8_t * values, + const float * quant_weights, + const int ntry) { + + float sigma2 = 0; + for (int j = 0; j < super_block_size; ++j) sigma2 += x[j]*x[j]; + sigma2 *= 2.f/super_block_size; + + memset(q4, 0, super_block_size/2); + dh[0] = GGML_FP32_TO_FP16(0.f); + + float max_scale = 0, amax_scale = 0; + for (int ib = 0; ib < super_block_size/block_size; ++ib) { + const float * xb = x + ib*block_size; + uint8_t * Lb = L + ib*block_size; + if (quant_weights) { + const float * qw = quant_weights + ib*block_size; + for (int j = 0; j < block_size; ++j) weight[j] = qw[j] * sqrtf(sigma2 + xb[j]*xb[j]); + } else { + for (int j = 0; j < block_size; ++j) weight[j] = xb[j]*xb[j]; + } + float amax = 0, max = 0; + for (int j = 0; j < block_size; ++j) { + float ax = fabsf(xb[j]); + if (ax > amax) { + amax = ax; max = xb[j]; + } + } + if (!amax) { + scales[ib] = 0; + continue; + } + float d = ntry > 0 ? -max/values[0] : max/values[0]; + float id = 1/d; + float sumqx = 0, sumq2 = 0; + for (int j = 0; j < block_size; ++j) { + float al = id*xb[j]; + int l = best_index_int8(16, values, al); + Lb[j] = l; + float q = values[l]; + float w = weight[j]; + sumqx += w*q*xb[j]; + sumq2 += w*q*q; + } + d = sumqx/sumq2; + float best = d*sumqx; + for (int itry = -ntry; itry <= ntry; ++itry) { + id = (itry + values[0])/max; + sumqx = sumq2 = 0; + for (int j = 0; j < block_size; ++j) { + float al = id*xb[j]; + int l = best_index_int8(16, values, al); + float q = values[l]; + float w = weight[j]; + sumqx += w*q*xb[j]; + sumq2 += w*q*q; + } + if (sumq2 > 0 && sumqx*sumqx > best*sumq2) { + d = sumqx/sumq2; best = d * sumqx; + } + } + scales[ib] = d; + float abs_d = fabsf(d); + if (abs_d > amax_scale) { + amax_scale = abs_d; max_scale = d; + } + } + + if (super_block_size/block_size > 1) { + int nb = super_block_size/block_size; + memset(scales_h, 0, ((nb+7)/8)*sizeof(uint16_t)); + float d = -max_scale/32; + dh[0] = GGML_FP32_TO_FP16(d); + float id = d ? 1/d : 0.f; + for (int ib = 0; ib < super_block_size/block_size; ++ib) { + int l = nearest_int(id*scales[ib]); + l = MAX(-32, MIN(31, l)); + float dl = d * l; + float idl = dl ? 1/dl : 0.f; + uint8_t * Lb = L + ib*block_size; + const float * xb = x + ib*block_size; + for (int j = 0; j < block_size; ++j) { + Lb[j] = best_index_int8(16, values, idl*xb[j]); + } + l += 32; + uint8_t l_l = l & 0xf; + uint8_t l_h = l >> 4; + if (ib%2 == 0) scales_l[ib/2] = l_l; + else scales_l[ib/2] |= (l_l << 4); + scales_h[ib/8] |= (l_h << 2*(ib%8)); + } + } else { + dh[0] = GGML_FP32_TO_FP16(scales[0]); + if (ntry > 0) { + float id = scales[0] ? 1/scales[0] : 0; + for (int j = 0; j < super_block_size; ++j) { + L[j] = best_index_int8(16, values, id*x[j]); + } + } + } - const float d_all = (float)x[i].d; + for (int i = 0; i < super_block_size/32; ++i) { + for (int j = 0; j < 16; ++j) { + q4[16*i + j] = L[32*i + j] | (L[32*i + 16 + j] << 4); + } + } +} - const uint8_t * restrict q6 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; +size_t quantize_iq4_nl(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK4_NL == 0); + int64_t nblock = n_per_row/QK4_NL; + char * qrow = (char *)dst; + uint8_t L[QK4_NL]; + float weight[QK4_NL]; + uint16_t unused_h; + uint8_t * unused_l = NULL; + float scale; + for (int64_t row = 0; row < nrow; ++row) { + block_iq4_nl * iq4 = (block_iq4_nl *)qrow; + for (int ibl = 0; ibl < nblock; ++ibl) { + const float * qw = quant_weights ? quant_weights + QK4_NL*ibl : NULL; + quantize_row_iq4_nl_impl(QK4_NL, 32, src + QK4_NL*ibl, &iq4[ibl].d, iq4[ibl].qs, &unused_h, unused_l, + &scale, weight, L, kvalues_iq4nl, qw, 7); + } + src += n_per_row; + qrow += nblock*sizeof(block_iq4_nl); + } + return nrow * nblock * sizeof(block_iq4_nl); +} - const int8_t * restrict scale = x[i].scales; +void quantize_row_iq4_nl(const float * restrict x, void * restrict vy, int64_t k) { + GGML_ASSERT(k%QK4_NL == 0); + int64_t nblock = k/QK4_NL; + uint8_t L[QK4_NL]; + float weight[QK4_NL]; + uint16_t unused_h; + uint8_t * unused_l = NULL; + float scale; + block_iq4_nl * iq4 = (block_iq4_nl *)vy; + for (int ibl = 0; ibl < nblock; ++ibl) { + quantize_row_iq4_nl_impl(QK4_NL, 32, x + QK4_NL*ibl, &iq4[ibl].d, iq4[ibl].qs, &unused_h, unused_l, + &scale, weight, L, kvalues_iq4nl, NULL, -1); + } +} - int32_t isum = 0; +void quantize_row_iq4_nl_reference(const float * restrict x, block_iq4_nl * restrict y, int64_t k) { + assert(k % QK4_NL == 0); + quantize_row_iq4_nl(x, y, k); +} - size_t vl = 16; +size_t quantize_iq4_xs(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { +#if QK_K == 64 + return quantize_iq4_nl(src, dst, nrow, n_per_row, quant_weights); +#else + GGML_ASSERT(n_per_row%QK_K == 0); + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + uint8_t L[QK_K]; + float weight[32]; + float scales[QK_K/32]; + for (int64_t row = 0; row < nrow; ++row) { + block_iq4_xs * iq4 = (block_iq4_xs *)qrow; + for (int ibl = 0; ibl < nblock; ++ibl) { + const float * qw = quant_weights ? quant_weights + QK_K*ibl : NULL; + quantize_row_iq4_nl_impl(QK_K, 32, src + QK_K*ibl, &iq4[ibl].d, iq4[ibl].qs, &iq4[ibl].scales_h, iq4[ibl].scales_l, + scales, weight, L, kvalues_iq4nl, qw, 7); + } + src += n_per_row; + qrow += nblock*sizeof(block_iq4_xs); + } + return nrow * nblock * sizeof(block_iq4_xs); +#endif +} - vint32m1_t vzero = __riscv_vmv_v_x_i32m1(0, 1); +void quantize_row_iq4_xs(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_iq4_xs * restrict y = vy; + quantize_row_iq4_xs_reference(x, y, k); +} - // load Q6 - vuint8mf2_t q6_0 = __riscv_vle8_v_u8mf2(q6, vl); - vuint8mf2_t q6_1 = __riscv_vle8_v_u8mf2(q6+16, vl); +void quantize_row_iq4_xs_reference(const float * restrict x, block_iq4_xs * restrict y, int64_t k) { + assert(k % QK_K == 0); + quantize_iq4_xs(x, y, 1, k, NULL); +} - // load qh - vuint8mf2_t qh_x = __riscv_vle8_v_u8mf2(qh, vl); +// =============================== 2.5625 bpw - vuint8mf2_t qh0 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); - qh_x = __riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl); - vuint8mf2_t qh1 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); - qh_x = __riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl); - vuint8mf2_t qh2 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); - qh_x = __riscv_vsrl_vx_u8mf2(qh_x, 0x2, vl); - vuint8mf2_t qh3 = __riscv_vsll_vx_u8mf2(__riscv_vand_vx_u8mf2(qh_x, 0x3, vl), 0x4, vl); +static void quantize_row_iq2_s_impl(const float * restrict x, void * restrict vy, int64_t n, const float * restrict quant_weights) { - vuint8mf2_t q6h_0 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(q6_0, 0xF, vl), qh0, vl); - vuint8mf2_t q6h_1 = __riscv_vor_vv_u8mf2(__riscv_vand_vx_u8mf2(q6_1, 0xF, vl), qh1, vl); - vuint8mf2_t q6h_2 = __riscv_vor_vv_u8mf2(__riscv_vsrl_vx_u8mf2(q6_0, 0x4, vl), qh2, vl); - vuint8mf2_t q6h_3 = __riscv_vor_vv_u8mf2(__riscv_vsrl_vx_u8mf2(q6_1, 0x4, vl), qh3, vl); + const int gindex = iq2_data_index(GGML_TYPE_IQ2_S); - vint8mf2_t q6v_0 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_0), 32, vl); - vint8mf2_t q6v_1 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_1), 32, vl); - vint8mf2_t q6v_2 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_2), 32, vl); - vint8mf2_t q6v_3 = __riscv_vsub_vx_i8mf2(__riscv_vreinterpret_v_u8mf2_i8mf2(q6h_3), 32, vl); + const uint64_t * kgrid_q2xs = iq2_data[gindex].grid; + const int * kmap_q2xs = iq2_data[gindex].map; + const uint16_t * kneighbors_q2xs = iq2_data[gindex].neighbours; - // load Q8 and take product - vint16m1_t p0 = __riscv_vwmul_vv_i16m1(q6v_0, __riscv_vle8_v_i8mf2(q8, vl), vl); - vint16m1_t p1 = __riscv_vwmul_vv_i16m1(q6v_1, __riscv_vle8_v_i8mf2(q8+16, vl), vl); - vint16m1_t p2 = __riscv_vwmul_vv_i16m1(q6v_2, __riscv_vle8_v_i8mf2(q8+32, vl), vl); - vint16m1_t p3 = __riscv_vwmul_vv_i16m1(q6v_3, __riscv_vle8_v_i8mf2(q8+48, vl), vl); + GGML_ASSERT(kmap_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kgrid_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(kneighbors_q2xs && "forgot to call ggml_quantize_init()?"); + GGML_ASSERT(n%QK_K == 0); - vint32m1_t vs_0 = __riscv_vwredsum_vs_i16m1_i32m1(p0, vzero, vl); - vint32m1_t vs_1 = __riscv_vwredsum_vs_i16m1_i32m1(p1, vzero, vl); - vint32m1_t vs_2 = __riscv_vwredsum_vs_i16m1_i32m1(p2, vzero, vl); - vint32m1_t vs_3 = __riscv_vwredsum_vs_i16m1_i32m1(p3, vzero, vl); + const int kMaxQ = 3; - isum += __riscv_vmv_x_s_i32m1_i32(vs_0) * scale[0]; - isum += __riscv_vmv_x_s_i32m1_i32(vs_1) * scale[1]; - isum += __riscv_vmv_x_s_i32m1_i32(vs_2) * scale[2]; - isum += __riscv_vmv_x_s_i32m1_i32(vs_3) * scale[3]; + const int64_t nbl = n/QK_K; - sumf += isum * d_all * y[i].d; + block_iq2_s * y = vy; - } + float scales[QK_K/16]; + float weight[16]; + float xval[16]; + int8_t L[16]; + int8_t Laux[16]; + float waux[16]; + bool is_on_grid[2]; + bool is_on_grid_aux[2]; + uint8_t block_signs[2]; - *s = sumf; + for (int ibl = 0; ibl < nbl; ++ibl) { -#else + memset(&y[ibl], 0, sizeof(block_iq2_s)); + y[ibl].d = GGML_FP32_TO_FP16(0.f); - int8_t aux8[QK_K]; - int16_t aux16[8]; - float sums [8]; - int32_t aux32[8]; - memset(sums, 0, 8*sizeof(float)); + float max_scale = 0; - float sumf = 0; - for (int i = 0; i < nb; ++i) { - const uint8_t * restrict q4 = x[i].ql; - const uint8_t * restrict qh = x[i].qh; - const int8_t * restrict q8 = y[i].qs; - memset(aux32, 0, 8*sizeof(int32_t)); - int8_t * restrict a = aux8; - for (int l = 0; l < 16; ++l) { - a[l+ 0] = (int8_t)((q4[l+ 0] & 0xF) | (((qh[l] >> 0) & 3) << 4)) - 32; - a[l+16] = (int8_t)((q4[l+16] & 0xF) | (((qh[l] >> 2) & 3) << 4)) - 32; - a[l+32] = (int8_t)((q4[l+ 0] >> 4) | (((qh[l] >> 4) & 3) << 4)) - 32; - a[l+48] = (int8_t)((q4[l+16] >> 4) | (((qh[l] >> 6) & 3) << 4)) - 32; + const float * xbl = x + QK_K*ibl; + float sumx2 = 0; + for (int i = 0; i < QK_K; ++i) sumx2 += xbl[i]*xbl[i]; + float sigma2 = 2*sumx2/QK_K; + + for (int ib = 0; ib < QK_K/16; ++ib) { + const float * xb = xbl + 16*ib; + if (quant_weights) { + const float * qw = quant_weights + QK_K*ibl + 16*ib; + for (int i = 0; i < 16; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]); + } else { + for (int i = 0; i < 16; ++i) weight[i] = 0.25f*sigma2 + xb[i]*xb[i]; + } + for (int i = 0; i < 16; ++i) waux[i] = sqrtf(weight[i]); + for (int k = 0; k < 2; ++k) { + uint8_t s = 0; + for (int i = 0; i < 8; ++i) { + if (xb[8*k + i] >= 0) xval[8*k + i] = xb[8*k + i]; + else { + xval[8*k + i] = -xb[8*k + i]; s |= (1 << i); + } + } + block_signs[k] = s; + } + float max = xval[0]; + for (int i = 1; i < 16; ++i) max = MAX(max, xval[i]); + if (!max) { + scales[ib] = 0; + continue; + } + float best = 0; + float scale = max/(2*kMaxQ-1); + is_on_grid[0] = is_on_grid[1] = true; + for (int is = -9; is <= 9; ++is) { + float id = (2*kMaxQ-1+is*0.1f)/max; + float this_scale = 1/id; + for (int k = 0; k < 2; ++k) { + for (int i = 0; i < 8; ++i) { + int l = nearest_int(0.5f*(id*xval[8*k+i]-1)); + Laux[8*k+i] = MAX(0, MIN(kMaxQ-1, l)); + } + uint16_t u = 0; + for (int i = 0; i < 8; ++i) u |= (Laux[8*k+i] << 2*i); + int grid_index = kmap_q2xs[u]; + is_on_grid_aux[k] = true; + if (grid_index < 0) { + is_on_grid_aux[k] = false; + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq2_find_best_neighbour(neighbours, kgrid_q2xs, xval + 8*k, waux + 8*k, this_scale, Laux + 8*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 16; ++i) { + float w = weight[i]; + float q = 2*Laux[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0 && sumqx*sumqx > best*sumq2) { + scale = sumqx/sumq2; best = scale*sumqx; + for (int i = 0; i < 16; ++i) L[i] = Laux[i]; + for (int k = 0; k < 2; ++k) is_on_grid[k] = is_on_grid_aux[k]; + } + } + int n_not_ongrid = 0; + for (int k = 0; k < 2; ++k) if (!is_on_grid[k]) ++n_not_ongrid; + if (n_not_ongrid > 0 && scale > 0) { + float id = 1/scale; + for (int k = 0; k < 2; ++k) { + if (is_on_grid[k]) continue; + uint16_t u = 0; + for (int i = 0; i < 8; ++i) { + int l = nearest_int(0.5f*(id*xval[8*k+i]-1)); + l = MAX(0, MIN(kMaxQ-1, l)); + u |= (l << 2*i); + L[8*k + i] = l; + } + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + const uint16_t * neighbours = kneighbors_q2xs - kmap_q2xs[u] - 1; + grid_index = iq2_find_best_neighbour(neighbours, kgrid_q2xs, xval + 8*k, waux + 8*k, scale, L + 8*k); + } + } + float sumqx = 0, sumq2 = 0; + for (int i = 0; i < 16; ++i) { + float w = weight[i]; + float q = 2*L[i] + 1; + sumqx += w*xval[i]*q; + sumq2 += w*q*q; + } + if (sumq2 > 0) scale = sumqx/sumq2; + } + if (scale < 0) { + scale = -scale; + for (int k = 0; k < 2; ++k) block_signs[k] = ~block_signs[k]; + } + for (int k = 0; k < 2; ++k) { + uint16_t u = 0; + for (int i = 0; i < 8; ++i) u |= (L[8*k+i] << 2*i); + int grid_index = kmap_q2xs[u]; + if (grid_index < 0) { + printf("Oops: found point %u not on grid:", u); + for (int i = 0; i < 8; ++i) printf(" %d", L[8*k+i]); + printf("\n"); + GGML_ASSERT(false); + } + const int i8 = 2*ib + k; + y[ibl].qs[i8] = grid_index & 255; + y[ibl].qh[i8/4] |= ((grid_index >> 8) << 2*(i8%4)); + y[ibl].qs[QK_K/8 + i8] = block_signs[k]; + } + GGML_ASSERT(scale >= 0); + scales[ib] = scale; + max_scale = MAX(max_scale, scale); } - int is = 0; - for (int j = 0; j < QK_K/16; ++j) { - int scale = x[i].scales[is++]; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; - for (int l = 0; l < 8; ++l) aux16[l] = q8[l] * a[l]; - for (int l = 0; l < 8; ++l) aux32[l] += scale * aux16[l]; - q8 += 8; a += 8; + + if (!max_scale) { + continue; + } + + float d = max_scale/31; + y[ibl].d = GGML_FP32_TO_FP16(d * 0.9875f); + float id = 1/d; + for (int ib = 0; ib < QK_K/16; ++ib) { + int l = nearest_int(0.5f*(id*scales[ib]-1)); + l = MAX(0, MIN(15, l)); + if (ib%2 == 0) y[ibl].scales[ib/2] = l; + else y[ibl].scales[ib/2] |= (l << 4); } - const float d = GGML_FP16_TO_FP32(x[i].d) * y[i].d; - for (int l = 0; l < 8; ++l) sums[l] += d * aux32[l]; } - for (int l = 0; l < 8; ++l) sumf += sums[l]; - *s = sumf; -#endif } -#endif +size_t quantize_iq2_s(const float * restrict src, void * restrict dst, int64_t nrow, int64_t n_per_row, const float * quant_weights) { + GGML_ASSERT(n_per_row%QK_K == 0); + int64_t nblock = n_per_row/QK_K; + char * qrow = (char *)dst; + for (int64_t row = 0; row < nrow; ++row) { + quantize_row_iq2_s_impl(src, qrow, n_per_row, quant_weights); + src += n_per_row; + qrow += nblock*sizeof(block_iq2_s); + } + return nrow * nblock * sizeof(block_iq2_s); +} + +void quantize_row_iq2_s_reference(const float * restrict x, block_iq2_s * restrict y, int64_t k) { + assert(k % QK_K == 0); + quantize_iq2_s(x, y, 1, k, NULL); +} + +void quantize_row_iq2_s(const float * restrict x, void * restrict vy, int64_t k) { + assert(k % QK_K == 0); + block_iq2_s * restrict y = vy; + quantize_row_iq2_s_reference(x, y, k); +} diff --git a/bindings/ruby/ext/ggml-quants.h b/bindings/ruby/ext/ggml-quants.h index 70c12c27465..4d436a8f06b 100644 --- a/bindings/ruby/ext/ggml-quants.h +++ b/bindings/ruby/ext/ggml-quants.h @@ -1,224 +1,133 @@ #pragma once -#include "ggml-impl.h" +#define GGML_COMMON_DECL_C +#include "ggml-common.h" -// GGML internal header - -#include -#include - -#define QK4_0 32 -typedef struct { - ggml_fp16_t d; // delta - uint8_t qs[QK4_0 / 2]; // nibbles / quants -} block_q4_0; -static_assert(sizeof(block_q4_0) == sizeof(ggml_fp16_t) + QK4_0 / 2, "wrong q4_0 block size/padding"); - -#define QK4_1 32 -typedef struct { - ggml_fp16_t d; // delta - ggml_fp16_t m; // min - uint8_t qs[QK4_1 / 2]; // nibbles / quants -} block_q4_1; -static_assert(sizeof(block_q4_1) == 2 * sizeof(ggml_fp16_t) + QK4_1 / 2, "wrong q4_1 block size/padding"); - -#define QK5_0 32 -typedef struct { - ggml_fp16_t d; // delta - uint8_t qh[4]; // 5-th bit of quants - uint8_t qs[QK5_0 / 2]; // nibbles / quants -} block_q5_0; -static_assert(sizeof(block_q5_0) == sizeof(ggml_fp16_t) + sizeof(uint32_t) + QK5_0 / 2, "wrong q5_0 block size/padding"); - -#define QK5_1 32 -typedef struct { - ggml_fp16_t d; // delta - ggml_fp16_t m; // min - uint8_t qh[4]; // 5-th bit of quants - uint8_t qs[QK5_1 / 2]; // nibbles / quants -} block_q5_1; -static_assert(sizeof(block_q5_1) == 2 * sizeof(ggml_fp16_t) + sizeof(uint32_t) + QK5_1 / 2, "wrong q5_1 block size/padding"); - -#define QK8_0 32 -typedef struct { - ggml_fp16_t d; // delta - int8_t qs[QK8_0]; // quants -} block_q8_0; -static_assert(sizeof(block_q8_0) == sizeof(ggml_fp16_t) + QK8_0, "wrong q8_0 block size/padding"); - -#define QK8_1 32 -typedef struct { - float d; // delta - float s; // d * sum(qs[i]) - int8_t qs[QK8_1]; // quants -} block_q8_1; -static_assert(sizeof(block_q8_1) == 2*sizeof(float) + QK8_1, "wrong q8_1 block size/padding"); - -// -// Super-block quantization structures -// - -// Super-block size -#ifdef GGML_QKK_64 -#define QK_K 64 -#define K_SCALE_SIZE 4 -#else -#define QK_K 256 -#define K_SCALE_SIZE 12 -#endif +#include "ggml.h" -// 2-bit quantization -// weight is represented as x = a * q + b -// 16 blocks of 16 elements each -// Effectively 2.5625 bits per weight -typedef struct { - uint8_t scales[QK_K/16]; // scales and mins, quantized with 4 bits - uint8_t qs[QK_K/4]; // quants - ggml_fp16_t d; // super-block scale for quantized scales - ggml_fp16_t dmin; // super-block scale for quantized mins -} block_q2_K; -static_assert(sizeof(block_q2_K) == 2*sizeof(ggml_fp16_t) + QK_K/16 + QK_K/4, "wrong q2_K block size/padding"); - -// 3-bit quantization -// weight is represented as x = a * q -// 16 blocks of 16 elements each -// Effectively 3.4375 bits per weight -#ifdef GGML_QKK_64 -typedef struct { - uint8_t hmask[QK_K/8]; // quants - high bit - uint8_t qs[QK_K/4]; // quants - low 2 bits - uint8_t scales[2]; - ggml_fp16_t d; // super-block scale -} block_q3_K; -static_assert(sizeof(block_q3_K) == sizeof(ggml_fp16_t) + QK_K / 4 + QK_K / 8 + 2, "wrong q3_K block size/padding"); -#else -typedef struct { - uint8_t hmask[QK_K/8]; // quants - high bit - uint8_t qs[QK_K/4]; // quants - low 2 bits - uint8_t scales[12]; // scales, quantized with 6 bits - ggml_fp16_t d; // super-block scale -} block_q3_K; -static_assert(sizeof(block_q3_K) == sizeof(ggml_fp16_t) + QK_K / 4 + QK_K / 8 + 12, "wrong q3_K block size/padding"); -#endif - -// 4-bit quantization -// 8 blocks of 32 elements each -// weight is represented as x = a * q + b -// Effectively 4.5 bits per weight -#ifdef GGML_QKK_64 -typedef struct { - ggml_fp16_t d[2]; // super-block scales/mins - uint8_t scales[2]; // 4-bit block scales/mins - uint8_t qs[QK_K/2]; // 4--bit quants -} block_q4_K; -static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + QK_K/2 + 2, "wrong q4_K block size/padding"); -#else -typedef struct { - ggml_fp16_t d; // super-block scale for quantized scales - ggml_fp16_t dmin; // super-block scale for quantized mins - uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits - uint8_t qs[QK_K/2]; // 4--bit quants -} block_q4_K; -static_assert(sizeof(block_q4_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/2, "wrong q4_K block size/padding"); -#endif +// GGML internal header -// 5-bit quantization -// 8 blocks of 32 elements each -// weight is represented as x = a * q + b -// Effectively 5.5 bits per weight -#ifdef GGML_QKK_64 -typedef struct { - ggml_fp16_t d; // super-block scale - int8_t scales[QK_K/16]; // 8-bit block scales - uint8_t qh[QK_K/8]; // quants, high bit - uint8_t qs[QK_K/2]; // quants, low 4 bits -} block_q5_K; -static_assert(sizeof(block_q5_K) == sizeof(ggml_fp16_t) + QK_K/2 + QK_K/8 + QK_K/16, "wrong q5_K block size/padding"); -#else -typedef struct { - ggml_fp16_t d; // super-block scale for quantized scales - ggml_fp16_t dmin; // super-block scale for quantized mins - uint8_t scales[K_SCALE_SIZE]; // scales and mins, quantized with 6 bits - uint8_t qh[QK_K/8]; // quants, high bit - uint8_t qs[QK_K/2]; // quants, low 4 bits -} block_q5_K; -static_assert(sizeof(block_q5_K) == 2*sizeof(ggml_fp16_t) + K_SCALE_SIZE + QK_K/2 + QK_K/8, "wrong q5_K block size/padding"); +#ifdef __cplusplus +extern "C" { #endif -// 6-bit quantization -// weight is represented as x = a * q -// 16 blocks of 16 elements each -// Effectively 6.5625 bits per weight -typedef struct { - uint8_t ql[QK_K/2]; // quants, lower 4 bits - uint8_t qh[QK_K/4]; // quants, upper 2 bits - int8_t scales[QK_K/16]; // scales, quantized with 8 bits - ggml_fp16_t d; // super-block scale -} block_q6_K; -static_assert(sizeof(block_q6_K) == sizeof(ggml_fp16_t) + QK_K / 16 + 3*QK_K/4, "wrong q6_K block size/padding"); - -// This is only used for intermediate quantization and dot products -typedef struct { - float d; // delta - int8_t qs[QK_K]; // quants - int16_t bsums[QK_K/16]; // sum of quants in groups of 16 -} block_q8_K; -static_assert(sizeof(block_q8_K) == sizeof(float) + QK_K + QK_K/16*sizeof(int16_t), "wrong q8_K block size/padding"); - - // Quantization -void quantize_row_q4_0_reference(const float * restrict x, block_q4_0 * restrict y, int k); -void quantize_row_q4_1_reference(const float * restrict x, block_q4_1 * restrict y, int k); -void quantize_row_q5_0_reference(const float * restrict x, block_q5_0 * restrict y, int k); -void quantize_row_q5_1_reference(const float * restrict x, block_q5_1 * restrict y, int k); -void quantize_row_q8_0_reference(const float * restrict x, block_q8_0 * restrict y, int k); -void quantize_row_q8_1_reference(const float * restrict x, block_q8_1 * restrict y, int k); - -void quantize_row_q2_K_reference(const float * restrict x, block_q2_K * restrict y, int k); -void quantize_row_q3_K_reference(const float * restrict x, block_q3_K * restrict y, int k); -void quantize_row_q4_K_reference(const float * restrict x, block_q4_K * restrict y, int k); -void quantize_row_q5_K_reference(const float * restrict x, block_q5_K * restrict y, int k); -void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict y, int k); -void quantize_row_q8_K_reference(const float * restrict x, block_q8_K * restrict y, int k); - -void quantize_row_q4_0(const float * restrict x, void * restrict y, int k); -void quantize_row_q4_1(const float * restrict x, void * restrict y, int k); -void quantize_row_q5_0(const float * restrict x, void * restrict y, int k); -void quantize_row_q5_1(const float * restrict x, void * restrict y, int k); -void quantize_row_q8_0(const float * restrict x, void * restrict y, int k); -void quantize_row_q8_1(const float * restrict x, void * restrict y, int k); - -void quantize_row_q2_K(const float * restrict x, void * restrict y, int k); -void quantize_row_q3_K(const float * restrict x, void * restrict y, int k); -void quantize_row_q4_K(const float * restrict x, void * restrict y, int k); -void quantize_row_q5_K(const float * restrict x, void * restrict y, int k); -void quantize_row_q6_K(const float * restrict x, void * restrict y, int k); -void quantize_row_q8_K(const float * restrict x, void * restrict y, int k); +void quantize_row_q4_0_reference(const float * GGML_RESTRICT x, block_q4_0 * GGML_RESTRICT y, int64_t k); +void quantize_row_q4_1_reference(const float * GGML_RESTRICT x, block_q4_1 * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_0_reference(const float * GGML_RESTRICT x, block_q5_0 * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_1_reference(const float * GGML_RESTRICT x, block_q5_1 * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_0_reference(const float * GGML_RESTRICT x, block_q8_0 * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_1_reference(const float * GGML_RESTRICT x, block_q8_1 * GGML_RESTRICT y, int64_t k); + +void quantize_row_q2_K_reference(const float * GGML_RESTRICT x, block_q2_K * GGML_RESTRICT y, int64_t k); +void quantize_row_q3_K_reference(const float * GGML_RESTRICT x, block_q3_K * GGML_RESTRICT y, int64_t k); +void quantize_row_q4_K_reference(const float * GGML_RESTRICT x, block_q4_K * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_K_reference(const float * GGML_RESTRICT x, block_q5_K * GGML_RESTRICT y, int64_t k); +void quantize_row_q6_K_reference(const float * GGML_RESTRICT x, block_q6_K * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_K_reference(const float * GGML_RESTRICT x, block_q8_K * GGML_RESTRICT y, int64_t k); + +void quantize_row_iq3_xxs_reference(const float * GGML_RESTRICT x, block_iq3_xxs * GGML_RESTRICT y, int64_t k); +void quantize_row_iq4_nl_reference (const float * GGML_RESTRICT x, block_iq4_nl * GGML_RESTRICT y, int64_t k); +void quantize_row_iq4_xs_reference (const float * GGML_RESTRICT x, block_iq4_xs * GGML_RESTRICT y, int64_t k); +void quantize_row_iq3_s_reference (const float * GGML_RESTRICT x, block_iq3_s * GGML_RESTRICT y, int64_t k); +void quantize_row_iq2_s_reference (const float * GGML_RESTRICT x, block_iq2_s * GGML_RESTRICT y, int64_t k); + +void quantize_row_q4_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q4_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_0(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_1(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); + +void quantize_row_q2_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q3_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q4_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q5_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q6_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_q8_K(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); + +void quantize_row_iq3_xxs(const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_iq4_nl (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_iq4_xs (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_iq3_s (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); +void quantize_row_iq2_s (const float * GGML_RESTRICT x, void * GGML_RESTRICT y, int64_t k); // Dequantization -void dequantize_row_q4_0(const block_q4_0 * restrict x, float * restrict y, int k); -void dequantize_row_q4_1(const block_q4_1 * restrict x, float * restrict y, int k); -void dequantize_row_q5_0(const block_q5_0 * restrict x, float * restrict y, int k); -void dequantize_row_q5_1(const block_q5_1 * restrict x, float * restrict y, int k); -void dequantize_row_q8_0(const block_q8_0 * restrict x, float * restrict y, int k); -//void dequantize_row_q8_1(const block_q8_1 * restrict x, float * restrict y, int k); - -void dequantize_row_q2_K(const block_q2_K * restrict x, float * restrict y, int k); -void dequantize_row_q3_K(const block_q3_K * restrict x, float * restrict y, int k); -void dequantize_row_q4_K(const block_q4_K * restrict x, float * restrict y, int k); -void dequantize_row_q5_K(const block_q5_K * restrict x, float * restrict y, int k); -void dequantize_row_q6_K(const block_q6_K * restrict x, float * restrict y, int k); -void dequantize_row_q8_K(const block_q8_K * restrict x, float * restrict y, int k); +void dequantize_row_q4_0(const block_q4_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q4_1(const block_q4_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q5_0(const block_q5_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q5_1(const block_q5_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q8_0(const block_q8_0 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +//void dequantize_row_q8_1(const block_q8_1 * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); + +void dequantize_row_q2_K(const block_q2_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q3_K(const block_q3_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q4_K(const block_q4_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q5_K(const block_q5_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q6_K(const block_q6_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_q8_K(const block_q8_K * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); + +void dequantize_row_iq2_xxs(const block_iq2_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq2_xs (const block_iq2_xs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq2_s (const block_iq2_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq3_xxs(const block_iq3_xxs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq1_s (const block_iq1_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq1_m (const block_iq1_m * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq4_nl (const block_iq4_nl * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq4_xs (const block_iq4_xs * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); +void dequantize_row_iq3_s (const block_iq3_s * GGML_RESTRICT x, float * GGML_RESTRICT y, int64_t k); // Dot product -void ggml_vec_dot_q4_0_q8_0(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q4_1_q8_1(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q5_0_q8_0(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q5_1_q8_1(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q8_0_q8_0(int n, float * restrict s, const void * restrict vx, const void * restrict vy); - -void ggml_vec_dot_q2_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q3_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q4_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q5_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); -void ggml_vec_dot_q6_K_q8_K(int n, float * restrict s, const void * restrict vx, const void * restrict vy); +void ggml_vec_dot_q4_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q4_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q5_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q5_1_q8_1(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q8_0_q8_0(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +void ggml_vec_dot_q2_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q3_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q4_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q5_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_q6_K_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +void ggml_vec_dot_iq2_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq2_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq2_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq3_xxs_q8_K(int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq1_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq1_m_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq4_nl_q8_0 (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq4_xs_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); +void ggml_vec_dot_iq3_s_q8_K (int n, float * GGML_RESTRICT s, size_t bs, const void * GGML_RESTRICT vx, size_t bx, const void * GGML_RESTRICT vy, size_t by, int nrc); + +// Quantization utilizing an importance matrix (a.k.a. "Activation aWare Quantization") +size_t quantize_iq2_xxs(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq2_xs (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq2_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq3_xxs(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq1_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq1_m (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq4_nl (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq4_xs (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_iq3_s (const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); + +size_t quantize_q2_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q3_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q4_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q5_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q6_K(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q4_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q4_1(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q5_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q5_1(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); +size_t quantize_q8_0(const float * GGML_RESTRICT src, void * GGML_RESTRICT dst, int64_t nrows, int64_t n_per_row, const float * imatrix); + +void iq2xs_init_impl(enum ggml_type type); +void iq2xs_free_impl(enum ggml_type type); +void iq3xs_init_impl(int grid_size); +void iq3xs_free_impl(int grid_size); + +#ifdef __cplusplus +} +#endif + diff --git a/bindings/ruby/ext/ggml-sycl.h b/bindings/ruby/ext/ggml-sycl.h new file mode 100644 index 00000000000..a9f776fc1dd --- /dev/null +++ b/bindings/ruby/ext/ggml-sycl.h @@ -0,0 +1,49 @@ +// +// MIT license +// Copyright (C) 2024 Intel Corporation +// SPDX-License-Identifier: MIT +// + +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#ifdef __cplusplus +extern "C" { +#endif + +#define GGML_SYCL_MAX_DEVICES 48 +#define GGML_SYCL_NAME "SYCL" + +// backend API +GGML_API ggml_backend_t ggml_backend_sycl_init(int device); + +// devide buffer +GGML_API ggml_backend_buffer_type_t ggml_backend_sycl_buffer_type(int device); + +// split tensor buffer that splits matrices by rows across multiple devices +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_sycl_split_buffer_type(const float * tensor_split); + +// pinned host buffer for use with the CPU backend for faster copies between CPU and GPU +GGML_API ggml_backend_buffer_type_t ggml_backend_sycl_host_buffer_type(void); + +GGML_API void ggml_backend_sycl_print_sycl_devices(void); +GGML_API GGML_CALL void ggml_sycl_get_gpu_list(int *id_list, int max_len); +GGML_API GGML_CALL void ggml_sycl_get_device_description(int device, char *description, size_t description_size); +GGML_API GGML_CALL int ggml_backend_sycl_get_device_count(); +GGML_API GGML_CALL void ggml_backend_sycl_get_device_memory(int device, size_t *free, size_t *total); +GGML_API GGML_CALL int ggml_backend_sycl_get_device_index(int device_id); + +// TODO: these are temporary +// ref: https://github.com/ggerganov/llama.cpp/pull/6022#issuecomment-1992615670 +GGML_API GGML_CALL int ggml_backend_sycl_get_device_id(int device_index); +GGML_API GGML_CALL void ggml_backend_sycl_set_single_device_mode(int main_gpu_id); +GGML_API GGML_CALL void ggml_backend_sycl_set_mul_device_mode(); + +// SYCL doesn't support registering host memory, keep here for reference +// GGML_API GGML_CALL bool ggml_backend_sycl_register_host_buffer(void * buffer, size_t size); +// GGML_API GGML_CALL void ggml_backend_sycl_unregister_host_buffer(void * buffer); +#ifdef __cplusplus +} +#endif diff --git a/bindings/ruby/ext/ggml-vulkan.h b/bindings/ruby/ext/ggml-vulkan.h new file mode 100644 index 00000000000..af661c2d7d5 --- /dev/null +++ b/bindings/ruby/ext/ggml-vulkan.h @@ -0,0 +1,29 @@ +#pragma once + +#include "ggml.h" +#include "ggml-backend.h" + +#ifdef __cplusplus +extern "C" { +#endif + +#define GGML_VK_NAME "Vulkan" +#define GGML_VK_MAX_DEVICES 16 + +GGML_API void ggml_vk_instance_init(void); + +// backend API +GGML_API GGML_CALL ggml_backend_t ggml_backend_vk_init(size_t dev_num); + +GGML_API GGML_CALL bool ggml_backend_is_vk(ggml_backend_t backend); +GGML_API GGML_CALL int ggml_backend_vk_get_device_count(void); +GGML_API GGML_CALL void ggml_backend_vk_get_device_description(int device, char * description, size_t description_size); +GGML_API GGML_CALL void ggml_backend_vk_get_device_memory(int device, size_t * free, size_t * total); + +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_vk_buffer_type(size_t dev_num); +// pinned host buffer for use with the CPU backend for faster copies between CPU and GPU +GGML_API GGML_CALL ggml_backend_buffer_type_t ggml_backend_vk_host_buffer_type(void); + +#ifdef __cplusplus +} +#endif diff --git a/bindings/ruby/whispercpp.gemspec b/bindings/ruby/whispercpp.gemspec new file mode 100644 index 00000000000..508a6a94052 --- /dev/null +++ b/bindings/ruby/whispercpp.gemspec @@ -0,0 +1,28 @@ +Gem::Specification.new do |s| + s.name = "whispercpp" + s.authors = ["Georgi Gerganov", "Todd A. Fisher"] + s.version = '1.3.0' + s.date = '2024-05-14' + s.description = %q{High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model via Ruby} + s.email = 'todd.fisher@gmail.com' + s.extra_rdoc_files = ['LICENSE', 'README.md'] + + s.files = ["LICENSE", "README.md", "Rakefile", "ext/extconf.rb", "ext/ggml.c", "ext/ruby_whisper.cpp", "ext/whisper.cpp", "ext/dr_wav.h", "ext/ggml.h", "ext/ruby_whisper.h", "ext/whisper.h"] + + #### Load-time details + s.require_paths = ['lib','ext'] + s.summary = %q{Ruby whisper.cpp bindings} + s.test_files = ["tests/test_whisper.rb"] + + s.extensions << 'ext/extconf.rb' + + + #### Documentation and testing. + s.homepage = 'https://github.com/ggerganov/whisper.cpp' + s.rdoc_options = ['--main', '../../README.md'] + + + s.platform = Gem::Platform::RUBY + + s.licenses = ['MIT'] +end From a7dc2aab16822b80a6491b0bd4bbf4900404a8a0 Mon Sep 17 00:00:00 2001 From: Daniel Valdivia <18384552+dvaldivia@users.noreply.github.com> Date: Sat, 25 May 2024 00:46:22 -0700 Subject: [PATCH 15/17] server : fix typo (#2181) A simple comment typo, PR can be dismissed --- examples/server/server.cpp | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/examples/server/server.cpp b/examples/server/server.cpp index c78b3026e18..2efa4c7a020 100644 --- a/examples/server/server.cpp +++ b/examples/server/server.cpp @@ -947,7 +947,7 @@ int main(int argc, char ** argv) { "application/json"); } - // reset params to thier defaults + // reset params to their defaults params = default_params; }); svr.Post(sparams.request_path + "/load", [&](const Request &req, Response &res){ From 05042a782db3e3df5e14dd992c72a89337648a53 Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Mon, 27 May 2024 10:20:25 +0300 Subject: [PATCH 16/17] Revert "whisper : remove extra backend instance (huh?)" (#2182) This reverts commit 4caa64b73ed4c0e71097c865b0f6a9c136b007c6. --- whisper.cpp | 19 +++++++++++++++---- 1 file changed, 15 insertions(+), 4 deletions(-) diff --git a/whisper.cpp b/whisper.cpp index 84aec8238cd..7b8c683fca7 100644 --- a/whisper.cpp +++ b/whisper.cpp @@ -818,6 +818,8 @@ struct whisper_state { whisper_decoder decoders[WHISPER_MAX_DECODERS]; + ggml_backend_t backend = nullptr; + // ggml-alloc: // - stores meta info about the intermediate tensors into the `meta` buffers // - stores the actual tensor data into the `data` buffers @@ -2261,7 +2263,7 @@ static bool whisper_encode_internal( } if (!whisper_encode_external(wstate)) { - if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { return false; } } else { @@ -2284,7 +2286,7 @@ static bool whisper_encode_internal( return false; } - if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { return false; } } @@ -2300,7 +2302,7 @@ static bool whisper_encode_internal( return false; } - if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { return false; } } @@ -2801,7 +2803,7 @@ static bool whisper_decode_internal( logits = gf->nodes[gf->n_nodes - 1]; - if (!ggml_graph_compute_helper(wctx.backend, gf, n_threads)) { + if (!ggml_graph_compute_helper(wstate.backend, gf, n_threads)) { return false; } } @@ -3248,6 +3250,13 @@ struct whisper_state * whisper_init_state(whisper_context * ctx) { whisper_state * state = new whisper_state; + state->backend = whisper_backend_init(ctx->params); + if (!state->backend) { + WHISPER_LOG_ERROR("%s: whisper_backend_init() failed\n", __func__); + whisper_free_state(state); + return nullptr; + } + // at this point, we don't know yet how many decoders will be used, so we overallocate 3x ctx // in theory, there can be a case where this is not enough, but in practice it should always be enough const int factor = 3; @@ -3684,6 +3693,8 @@ void whisper_free_state(struct whisper_state * state) { ggml_gallocr_free(state->alloc_cross.alloc); ggml_gallocr_free(state->alloc_decode.alloc); + ggml_backend_free(state->backend); + // [EXPERIMENTAL] Token-level timestamps with DTW aheads_masks_free(state->aheads_masks); From c7b6988678779901d02ceba1a8212d2c9908956e Mon Sep 17 00:00:00 2001 From: Georgi Gerganov Date: Mon, 27 May 2024 10:35:09 +0300 Subject: [PATCH 17/17] release : v1.6.2 --- CMakeLists.txt | 2 +- README.md | 2 +- bindings/ios | 2 +- bindings/javascript/package.json | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) diff --git a/CMakeLists.txt b/CMakeLists.txt index 541be8a5d57..82913aa62ba 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -3,7 +3,7 @@ cmake_minimum_required (VERSION 3.5) # Allow for the creation of solution folders. set_property(GLOBAL PROPERTY USE_FOLDERS ON) -project(whisper.cpp VERSION 1.6.1) +project(whisper.cpp VERSION 1.6.2) set(SOVERSION 1) # Add path to modules diff --git a/README.md b/README.md index 0c34e8dffce..c2c6bc4a2b1 100644 --- a/README.md +++ b/README.md @@ -6,7 +6,7 @@ [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT) [![npm](https://img.shields.io/npm/v/whisper.cpp.svg)](https://www.npmjs.com/package/whisper.cpp/) -Stable: [v1.6.0](https://github.com/ggerganov/whisper.cpp/releases/tag/v1.6.0) / [Roadmap | F.A.Q.](https://github.com/ggerganov/whisper.cpp/discussions/126) +Stable: [v1.6.2](https://github.com/ggerganov/whisper.cpp/releases/tag/v1.6.0) / [Roadmap | F.A.Q.](https://github.com/ggerganov/whisper.cpp/discussions/126) High-performance inference of [OpenAI's Whisper](https://github.com/openai/whisper) automatic speech recognition (ASR) model: diff --git a/bindings/ios b/bindings/ios index 9a32de38144..a2085436c2e 160000 --- a/bindings/ios +++ b/bindings/ios @@ -1 +1 @@ -Subproject commit 9a32de3814477ad2e598d4a550fcab4b23a9c576 +Subproject commit a2085436c2eb796af90956b62bd64731f5e5b823 diff --git a/bindings/javascript/package.json b/bindings/javascript/package.json index da6a9efdc6c..2b3c806f353 100644 --- a/bindings/javascript/package.json +++ b/bindings/javascript/package.json @@ -1,6 +1,6 @@ { "name": "whisper.cpp", - "version": "1.6.1", + "version": "1.6.2", "description": "Whisper speech recognition", "main": "whisper.js", "scripts": {