traveller59
Cannot build generated code with cuda13
When trying to compile generated code with cuda 13, cuda complains it requires CUDA_STANDARD 17 (C++17) to be enabled, but the generated code forces CUDA_STANDARD 14 (c++14).
I would probably be a good idea not tu use the C++ version in the code but use compile_features instead.
Hey were you able to compile with cuda 13 ? I am also trying to getting nvcc fatal : Unsupported gpu architecture 'compute_52' not sure where 52 arch is being set. Mine is sm_120
I was able to compile by changing CXX_STANDARD 14 CUDA_STANDARD 14 to CXX_STANDARD 17 CUDA_STANDARD 17
@Arthav24 Where did you make those changes ? I have been trying to get it to build with cuda 13.0 by setting std in setup file to c++17 but it doesn't work. Thanks !
find . -type f -exec sed -i -e 's/CXX_STANDARD 14/CXX_STANDARD 17/g' -e 's/CUDA_STANDARD 14/CUDA_STANDARD 17/g' {} +
