[Kernel] Triton implementation of causal-conv1d for Mamba-based models (#18218)

Signed-off-by: Tuan M. Hoang-Trong <tmhoangt@us.ibm.com>
Co-authored-by: Tuan M. Hoang-Trong <tmhoangt@us.ibm.com>
Co-authored-by: Tyler Michael Smith <tysmith@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>

Author: 3 people

CMakeLists.txt

@@ -232,7 +232,6 @@ endif()
 
 set(VLLM_EXT_SRC
   "csrc/mamba/mamba_ssm/selective_scan_fwd.cu"
-  "csrc/mamba/causal_conv1d/causal_conv1d.cu"
   "csrc/cache_kernels.cu"
   "csrc/attention/paged_attention_v1.cu"
   "csrc/attention/paged_attention_v2.cu"

csrc/mamba/causal_conv1d/causal_conv1d.cu

@@ -326,22 +326,6 @@ void selective_scan_fwd(const torch::Tensor& u, const torch::Tensor& delta,
                         const std::optional<torch::Tensor>& has_initial_state,
                         const torch::Tensor& ssm_states, int64_t pad_slot_id);
 
-void causal_conv1d_update(const at::Tensor& x, const at::Tensor& conv_state,
-                          const at::Tensor& weight,
-                          const std::optional<at::Tensor>& bias_,
-                          bool silu_activation,
-                          const std::optional<at::Tensor>& cache_seqlens_,
-                          const std::optional<at::Tensor>& conv_state_indices_,
-                          int64_t pad_slot_id);
-
-void causal_conv1d_fwd(const at::Tensor& x, const at::Tensor& weight,
-                       const std::optional<at::Tensor>& bias_,
-                       const std::optional<at::Tensor>& conv_states,
-                       const std::optional<at::Tensor>& query_start_loc,
-                       const std::optional<at::Tensor>& cache_indices,
-                       const std::optional<at::Tensor>& has_initial_state,
-                       bool silu_activation, int64_t pad_slot_id);
-
 using fptr_t = int64_t;
 fptr_t init_custom_ar(const std::vector<int64_t>& fake_ipc_ptrs,
                       torch::Tensor& rank_data, int64_t rank,

csrc/mamba/causal_conv1d/causal_conv1d.h

@@ -594,28 +594,6 @@ TORCH_LIBRARY_EXPAND(TORCH_EXTENSION_NAME, ops) {
       "int pad_slot_id) -> ()");
   ops.impl("selective_scan_fwd", torch::kCUDA, &selective_scan_fwd);
 
-  ops.def(
-      "causal_conv1d_update(Tensor! x,"
-      "Tensor! conv_state,"
-      "Tensor! weight,"
-      "Tensor? bias_,"
-      "bool silu_activation,"
-      "Tensor? cache_seqlens_,"
-      "Tensor? conv_state_indices,"
-      "int pad_slot_id) -> ()");
-  ops.impl("causal_conv1d_update", torch::kCUDA, &causal_conv1d_update);
-
-  ops.def(
-      "causal_conv1d_fwd(Tensor! x, Tensor! weight,"
-      "Tensor? bias_,"
-      "Tensor!? conv_states,"
-      "Tensor? query_start_loc,"
-      "Tensor? cache_indices,"
-      "Tensor? has_initial_state,"
-      "bool silu_activation,"
-      "int pad_slot_id) -> ()");
-  ops.impl("causal_conv1d_fwd", torch::kCUDA, &causal_conv1d_fwd);
-
 #ifndef USE_ROCM
   // reorder weight for AllSpark Ampere W8A16 Fused Gemm kernel
   ops.def(