happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

cuda: fix compile error for cuda backend 

If we compile cuda backend, we must set different
cuda nvcc path to CMAKE_CUDA_COMPILER for diff
operating system, it's too troublesome.

Signed-off-by: lizhenneng <lizhenneng@kylinos.cn>
This commit is contained in:
lizhenneng 2025-12-22 10:58:04 +08:00
parent e1f15b454f
commit 02cb51df50
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
ggml/src/ggml-cuda/CMakeLists.txt
View File

@ -38,6 +38,13 @@ if (CUDAToolkit_FOUND)
endif() endif()
message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}") message(STATUS "Using CUDA architectures: ${CMAKE_CUDA_ARCHITECTURES}")
if(NOT DEFINED CMAKE_CUDA_COMPILER)
if(WIN32)
set(CMAKE_CUDA_COMPILER "${CUDAToolkit_BIN_DIR}/nvcc.exe")
else()
set(CMAKE_CUDA_COMPILER "${CUDAToolkit_BIN_DIR}/nvcc")
endif()
endif()
enable_language(CUDA) enable_language(CUDA)
file(GLOB GGML_HEADERS_CUDA "*.cuh") file(GLOB GGML_HEADERS_CUDA "*.cuh")