tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-07-05 11:13:30 +00:00
Code Issues Packages Projects Releases Wiki Activity

ggml : add GGML_CUDA_USE_GRAPHS option, restore GGML_CUDA_FORCE_CUBLAS (cmake) (#8140)

Browse Source
This commit is contained in:
slaren
2024-06-26 21:34:14 +02:00
committed by GitHub
parent c7ab7b612c
commit 31ec3993f6
3 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
ggml/src/CMakeLists.txt
View File

@ -295,12 +295,15 @@ if (GGML_CUDA)
list(APPEND GGML_CDEF_PUBLIC GGML_USE_CUDA)
add_compile_definitions(GGML_CUDA_USE_GRAPHS)
add_compile_definitions(GGML_CUDA_DMMV_X=${GGML_CUDA_DMMV_X})
add_compile_definitions(GGML_CUDA_MMV_Y=${GGML_CUDA_MMV_Y})
add_compile_definitions(K_QUANTS_PER_ITERATION=${GGML_CUDA_KQUANTS_ITER})
add_compile_definitions(GGML_CUDA_PEER_MAX_BATCH_SIZE=${GGML_CUDA_PEER_MAX_BATCH_SIZE})
if (GGML_CUDA_USE_GRAPHS)
add_compile_definitions(GGML_CUDA_USE_GRAPHS)
endif()
if (GGML_CUDA_FORCE_DMMV)
add_compile_definitions(GGML_CUDA_FORCE_DMMV)
endif()