vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

fix cmake

Browse source
This commit is contained in:
Johannes Gäßler 2024-05-29 20:10:10 +02:00
parent af95ae49a3
commit 9740ae0adc
1 changed files with 26 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

26
CMakeLists.txt
View file

@ -403,6 +403,8 @@ if (LLAMA_CUDA)
file(GLOB GGML_SOURCES_CUDA "ggml-cuda/*.cu") file(GLOB GGML_SOURCES_CUDA "ggml-cuda/*.cu")
list(APPEND GGML_SOURCES_CUDA "ggml-cuda.cu") list(APPEND GGML_SOURCES_CUDA "ggml-cuda.cu")
file(GLOB SRCS "ggml-cuda/template-instances/fattn-wmma*.cu")
list(APPEND GGML_SOURCES_CUDA ${SRCS})
add_compile_definitions(GGML_USE_CUDA) add_compile_definitions(GGML_USE_CUDA)
add_compile_definitions(GGML_CUDA_USE_GRAPHS) add_compile_definitions(GGML_CUDA_USE_GRAPHS)
@ -429,7 +431,16 @@ if (LLAMA_CUDA)
add_compile_definitions(GGML_CUDA_NO_PEER_COPY) add_compile_definitions(GGML_CUDA_NO_PEER_COPY)
endif() endif()
if (LLAMA_CUDA_FA_ALL_QUANTS) if (LLAMA_CUDA_FA_ALL_QUANTS)
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*.cu")
list(APPEND GGML_SOURCES_CUDA ${SRCS})
add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS) add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS)
else()
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu")
list(APPEND GGML_SOURCES_CUDA ${SRCS})
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu")
list(APPEND GGML_SOURCES_CUDA ${SRCS})
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*f16-f16.cu")
list(APPEND GGML_SOURCES_CUDA ${SRCS})
endif() endif()
if (LLAMA_STATIC) if (LLAMA_STATIC)
@ -575,6 +586,8 @@ if (LLAMA_HIPBLAS)
file(GLOB GGML_SOURCES_ROCM "ggml-cuda/*.cu") file(GLOB GGML_SOURCES_ROCM "ggml-cuda/*.cu")
list(APPEND GGML_SOURCES_ROCM "ggml-cuda.cu") list(APPEND GGML_SOURCES_ROCM "ggml-cuda.cu")
file(GLOB SRCS "ggml-cuda/template-instances/fattn-wmma*.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
add_compile_definitions(GGML_USE_HIPBLAS GGML_USE_CUDA) add_compile_definitions(GGML_USE_HIPBLAS GGML_USE_CUDA)
@ -594,6 +607,19 @@ if (LLAMA_HIPBLAS)
add_compile_definitions(GGML_CUDA_NO_PEER_COPY) add_compile_definitions(GGML_CUDA_NO_PEER_COPY)
endif() endif()
if (LLAMA_CUDA_FA_ALL_QUANTS)
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
add_compile_definitions(GGML_CUDA_FA_ALL_QUANTS)
else()
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q4_0-q4_0.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*q8_0-q8_0.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
file(GLOB SRCS "ggml-cuda/template-instances/fattn-vec*f16-f16.cu")
list(APPEND GGML_SOURCES_ROCM ${SRCS})
endif()
add_compile_definitions(GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X}) add_compile_definitions(GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X})
add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_MMV_Y}) add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_MMV_Y})
add_compile_definitions(K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER}) add_compile_definitions(K_QUANTS_PER_ITERATION=${LLAMA_CUDA_KQUANTS_ITER})