tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-26 19:55:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add an option to build without CUDA VMM (#7067)

Browse Source 
Add an option to build ggml cuda without CUDA VMM
resolves
https://github.com/ggerganov/llama.cpp/issues/6889
https://forums.developer.nvidia.com/t/potential-nvshmem-allocated-memory-performance-issue/275416/4
This commit is contained in:
William Tambellini
2024-05-06 11:12:14 -07:00
committed by GitHub
parent b3a995b416
commit 858f6b73f6
2 changed files with 13 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
CMakeLists.txt
View File

@ -103,6 +103,8 @@ set(LLAMA_CUDA_KQUANTS_ITER "2" CACHE STRING "llama: iters./thread per block for
set(LLAMA_CUDA_PEER_MAX_BATCH_SIZE "128" CACHE STRING
"llama: max. batch size for using peer access")
option(LLAMA_CUDA_NO_PEER_COPY "llama: do not use peer to peer copies" OFF)
option(LLAMA_CUDA_NO_VMM "llama: do not try to use CUDA VMM" OFF)
option(LLAMA_CURL "llama: use libcurl to download model from an URL" OFF)
option(LLAMA_HIPBLAS "llama: use hipBLAS" OFF)
option(LLAMA_HIP_UMA "llama: use HIP unified memory architecture" OFF)
@ -409,6 +411,9 @@ if (LLAMA_CUDA)
if (LLAMA_CUDA_FORCE_MMQ)
add_compile_definitions(GGML_CUDA_FORCE_MMQ)
endif()
if (LLAMA_CUDA_NO_VMM)
add_compile_definitions(GGML_CUDA_NO_VMM)
endif()
add_compile_definitions(GGML_CUDA_DMMV_X=${LLAMA_CUDA_DMMV_X})
add_compile_definitions(GGML_CUDA_MMV_Y=${LLAMA_CUDA_MMV_Y})
if (DEFINED LLAMA_CUDA_DMMV_Y)
@ -434,7 +439,11 @@ if (LLAMA_CUDA)
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} CUDA::cudart CUDA::cublas CUDA::cublasLt)
endif()
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} CUDA::cuda_driver)
if (LLAMA_CUDA_NO_VMM)
# No VMM requested, no need to link directly with the cuda driver lib (libcuda.so)
else()
set(LLAMA_EXTRA_LIBS ${LLAMA_EXTRA_LIBS} CUDA::cuda_driver) # required by cuDeviceGetAttribute(), cuMemGetAllocationGranularity(...), ...
endif()
if (NOT DEFINED CMAKE_CUDA_ARCHITECTURES)
# 52 == lowest CUDA 12 standard