pytorch/cmake/Modules_CUDA_fix/upstream/FindCUDA
eqy 9963845a4e [CUDA] Support family-conditional compute capabilies in TORCH_CUDA_ARCH_LIST (#157999)
Similar to arch-conditionals, such as 9.0a  and 10.0a, family conditionals such as 10.0f enable features specific to a family of architectures, such as between sm100 and sm103

Pull Request resolved: https://github.com/pytorch/pytorch/pull/157999
Approved by: https://github.com/Skylion007

Co-authored-by: Aaron Gokaslan <aaronGokaslan@gmail.com>
2025-07-11 20:34:59 +00:00
..
make2cmake.cmake Revert "Use official CUDAToolkit module in CMake (#154595)" 2025-06-23 21:15:31 +00:00
parse_cubin.cmake Revert "Use official CUDAToolkit module in CMake (#154595)" 2025-06-23 21:15:31 +00:00
run_nvcc.cmake Revert "Use official CUDAToolkit module in CMake (#154595)" 2025-06-23 21:15:31 +00:00
select_compute_arch.cmake [CUDA] Support family-conditional compute capabilies in TORCH_CUDA_ARCH_LIST (#157999) 2025-07-11 20:34:59 +00:00