ggml : remove kompute backend (#14501)

ggml-ci

Author: ggerganov

.github/ISSUE_TEMPLATE/010-bug-compilation.yml

@@ -40,7 +40,7 @@ body:
     attributes:
         label: GGML backends
         description: Which GGML backends do you know to be affected?
-        options: [AMX, BLAS, CPU, CUDA, HIP, Kompute, Metal, Musa, RPC, SYCL, Vulkan, OpenCL]
+        options: [AMX, BLAS, CPU, CUDA, HIP, Metal, Musa, RPC, SYCL, Vulkan, OpenCL]
         multiple: true
     validations:
       required: true

.github/ISSUE_TEMPLATE/011-bug-results.yml

@@ -42,7 +42,7 @@ body:
     attributes:
         label: GGML backends
         description: Which GGML backends do you know to be affected?
-        options: [AMX, BLAS, CPU, CUDA, HIP, Kompute, Metal, Musa, RPC, SYCL, Vulkan, OpenCL]
+        options: [AMX, BLAS, CPU, CUDA, HIP, Metal, Musa, RPC, SYCL, Vulkan, OpenCL]
         multiple: true
     validations:
       required: true

.github/labeler.yml

@@ -1,10 +1,4 @@
 # https://github.com/actions/labeler
-Kompute:
-    - changed-files:
-        - any-glob-to-any-file:
-            - ggml/include/ggml-kompute.h
-            - ggml/src/ggml-kompute/**
-            - README-kompute.md
 Apple Metal:
     - changed-files:
         - any-glob-to-any-file:

.github/workflows/build.yml

@@ -740,9 +740,6 @@ jobs:
           - build: 'llvm-arm64-opencl-adreno'
             arch: 'arm64'
             defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/arm64-windows-llvm.cmake -DCMAKE_PREFIX_PATH="$env:RUNNER_TEMP/opencl-arm64-release" -DGGML_OPENCL=ON -DGGML_OPENCL_USE_ADRENO_KERNELS=ON'
-         # - build: 'kompute-x64'
-         #   arch: 'x64'
-         #   defines: '-G "Ninja Multi-Config" -D CMAKE_TOOLCHAIN_FILE=cmake/x64-windows-llvm.cmake -DGGML_NATIVE=OFF -DLLAMA_BUILD_SERVER=ON -DGGML_RPC=ON -DGGML_BACKEND_DL=ON -DGGML_CPU_ALL_VARIANTS=ON -DGGML_OPENMP=OFF -DGGML_KOMPUTE=ON -DKOMPUTE_OPT_DISABLE_VULKAN_VERSION_CHECK=ON'
 
     steps:
       - name: Clone
@@ -756,12 +753,6 @@ jobs:
           variant: ccache
           evict-old-files: 1d
 
-      - name: Clone Kompute submodule
-        id: clone_kompute
-        if: ${{ matrix.build == 'kompute-x64' }}
-        run: |
-          git submodule update --init ggml/src/ggml-kompute/kompute
-
       - name: Download OpenBLAS
         id: get_openblas
         if: ${{ matrix.build == 'openblas-x64' }}
@@ -777,7 +768,7 @@ jobs:
 
       - name: Install Vulkan SDK
         id: get_vulkan
-        if: ${{ matrix.build == 'kompute-x64' || matrix.build == 'vulkan-x64' }}
+        if: ${{ matrix.build == 'vulkan-x64' }}
         run: |
           curl.exe -o $env:RUNNER_TEMP/VulkanSDK-Installer.exe -L "https://sdk.lunarg.com/sdk/download/${env:VULKAN_VERSION}/windows/vulkansdk-windows-X64-${env:VULKAN_VERSION}.exe"
           & "$env:RUNNER_TEMP\VulkanSDK-Installer.exe" --accept-licenses --default-answer --confirm-command install

.gitmodules

@@ -1,3 +0,0 @@
-[submodule "kompute"]
-	path = ggml/src/ggml-kompute/kompute
-	url = https://github.com/nomic-ai/kompute.git

CMakeLists.txt

@@ -120,7 +120,6 @@ endfunction()
 
 llama_option_depr(FATAL_ERROR LLAMA_CUBLAS              GGML_CUDA)
 llama_option_depr(WARNING     LLAMA_CUDA                GGML_CUDA)
-llama_option_depr(WARNING     LLAMA_KOMPUTE             GGML_KOMPUTE)
 llama_option_depr(WARNING     LLAMA_METAL               GGML_METAL)
 llama_option_depr(WARNING     LLAMA_METAL_EMBED_LIBRARY GGML_METAL_EMBED_LIBRARY)
 llama_option_depr(WARNING     LLAMA_NATIVE              GGML_NATIVE)

ggml/CMakeLists.txt

@@ -181,7 +181,6 @@ option(GGML_VULKAN_MEMORY_DEBUG             "ggml: enable Vulkan memory debug ou
 option(GGML_VULKAN_SHADER_DEBUG_INFO        "ggml: enable Vulkan shader debug info"           OFF)
 option(GGML_VULKAN_VALIDATE                 "ggml: enable Vulkan validation"                  OFF)
 option(GGML_VULKAN_RUN_TESTS                "ggml: run Vulkan tests"                          OFF)
-option(GGML_KOMPUTE                         "ggml: use Kompute"                               OFF)
 option(GGML_METAL                           "ggml: use Metal"                                 ${GGML_METAL_DEFAULT})
 option(GGML_METAL_USE_BF16                  "ggml: use bfloat if available"                   OFF)
 option(GGML_METAL_NDEBUG                    "ggml: disable Metal debugging"                   OFF)
@@ -266,7 +265,6 @@ set(GGML_PUBLIC_HEADERS
     include/ggml-cann.h
     include/ggml-cpp.h
     include/ggml-cuda.h
-    include/ggml-kompute.h
     include/ggml-opt.h
     include/ggml-metal.h
     include/ggml-rpc.h

ggml/include/ggml-kompute.h

@@ -365,7 +365,6 @@ ggml_add_backend(BLAS)
 ggml_add_backend(CANN)
 ggml_add_backend(CUDA)
 ggml_add_backend(HIP)
-ggml_add_backend(Kompute)
 ggml_add_backend(METAL)
 ggml_add_backend(MUSA)
 ggml_add_backend(RPC)

ggml/src/CMakeLists.txt

@@ -61,10 +61,6 @@
 #include "ggml-cann.h"
 #endif
 
-#ifdef GGML_USE_KOMPUTE
-#include "ggml-kompute.h"
-#endif
-
 // disable C++17 deprecation warning for std::codecvt_utf8
 #if defined(__clang__)
 #    pragma clang diagnostic push
@@ -189,9 +185,6 @@ struct ggml_backend_registry {
 #ifdef GGML_USE_RPC
         register_backend(ggml_backend_rpc_reg());
 #endif
-#ifdef GGML_USE_KOMPUTE
-        register_backend(ggml_backend_kompute_reg());
-#endif
 #ifdef GGML_USE_CPU
         register_backend(ggml_backend_cpu_reg());
 #endif
@@ -575,7 +568,6 @@ void ggml_backend_load_all_from_path(const char * dir_path) {
     ggml_backend_load_best("cann", silent, dir_path);
     ggml_backend_load_best("cuda", silent, dir_path);
     ggml_backend_load_best("hip", silent, dir_path);
-    ggml_backend_load_best("kompute", silent, dir_path);
     ggml_backend_load_best("metal", silent, dir_path);
     ggml_backend_load_best("rpc", silent, dir_path);
     ggml_backend_load_best("sycl", silent, dir_path);