Arthur-Ling commented 11 months ago

When I run "bash build_cutlass.sh", an error happens:

-- CMake Version: 3.18.2 -- The CXX compiler identification is GNU 8.4.1 -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Check for working CXX compiler: /usr/bin/g++ - skipped -- Detecting CXX compile features -- Detecting CXX compile features - done CMake Error at /usr/share/cmake/Modules/CMakeDetermineCUDACompiler.cmake:25 (message): Could not find compiler set in environment variable CUDACXX:

/usr/local/cuda/bin/nvcc.

Call Stack (most recent call first): CUDA.cmake:46 (enable_language) CMakeLists.txt:42 (include)

CMake Error: CMAKE_CUDA_COMPILER not set, after EnableLanguage -- Configuring incomplete, errors occurred! See also "/home/lizhangming/Project/torch-int-main/submodules/cutlass/build/CMakeFiles/CMakeOutput.log". make: *** No targets specified and no makefile found. Stop.

what should I do before "bash build_cutlass.sh"?

DonliFly commented 8 months ago

1) you should check the path /usr/local/cuda/bin/nvcc, maybe the path not exist; 2) then find the nvcc path, My path is /root/miniconda3/bin/nvcc 3) modified the build_cutlass.sh file, like this

export CUDACXX=/usr/local/cuda/bin/nvcc

export CUDACXX=/root/miniconda3/bin/nvcc

then it will work

awer-A commented 8 months ago

@DonliFly Hi，How can I find the path to nvcc?

Guangxuan-Xiao / torch-int

Could not find compiler set in environment variable CUDACXX #25

export CUDACXX=/usr/local/cuda/bin/nvcc