Skip to content

Fix handling of the 9.0a and 12.0a capabilities #202

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Aug 29, 2025
Merged

Fix handling of the 9.0a and 12.0a capabilities #202

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
14 changes: 1 addition & 13 deletions build2cmake/src/cuda_supported_archs.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,13 +1 @@
[
"7.0",
"7.2",
"7.5",
"8.0",
"8.6",
"8.7",
"8.9",
"9.0+PTX",
"10.0",
"10.1",
"12.0+PTX"
]
["7.0", "7.2", "7.5", "8.0", "8.6", "8.7", "8.9", "9.0", "10.0", "10.1", "12.0"]
2 changes: 1 addition & 1 deletion build2cmake/src/templates/cuda/kernel.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,7 +18,7 @@ if(GPU_LANG STREQUAL "CUDA")
{% if cuda_capabilities %}
cuda_archs_loose_intersection({{kernel_name}}_ARCHS "{{ cuda_capabilities|join(";") }}" "${CUDA_ARCHS}")
{% else %}
cuda_archs_loose_intersection({{kernel_name}}_ARCHS "${CUDA_SUPPORTED_ARCHS}" "${CUDA_ARCHS}")
cuda_archs_loose_intersection({{kernel_name}}_ARCHS "${CUDA_DEFAULT_KERNEL_ARCHS}" "${CUDA_ARCHS}")
{% endif %}
message(STATUS "Capabilities for kernel {{kernel_name}}: {{ '${' + kernel_name + '_ARCHS}'}}")
set_gencode_flags_for_srcs(SRCS {{'"${' + kernel_name + '_SRC}"'}} CUDA_ARCHS "{{ '${' + kernel_name + '_ARCHS}'}}")
Expand Down
7 changes: 7 additions & 0 deletions build2cmake/src/templates/cuda/preamble.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,13 @@ if (NOT TARGET_DEVICE STREQUAL "cuda" AND
return()
endif()

if(DEFINED CMAKE_CUDA_COMPILER_VERSION AND
CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 12.8)
set(CUDA_DEFAULT_KERNEL_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0;10.0;10.1;12.0+PTX")
else()
set(CUDA_DEFAULT_KERNEL_ARCHS "7.0;7.2;7.5;8.0;8.6;8.7;8.9;9.0+PTX")
endif()

if (NOT HIP_FOUND AND CUDA_FOUND)
set(GPU_LANG "CUDA")

Expand Down
4 changes: 2 additions & 2 deletions lib/torch-extension/default.nix
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -132,9 +132,9 @@ stdenv.mkDerivation (prevAttrs: {
CUDAToolkit_ROOT = "${lib.getDev cudaPackages.cuda_nvcc}";
TORCH_CUDA_ARCH_LIST =
if cudaPackages.cudaOlder "12.8" then
"7.0;7.5;8.0;8.6;8.9;9.0+PTX"
"7.0;7.5;8.0;8.6;8.9;9.0"
else
"7.0;7.5;8.0;8.6;8.9;9.0;10.0;10.1;12.0+PTX";
"7.0;7.5;8.0;8.6;8.9;9.0;10.0;10.1;12.0";
}
// lib.optionalAttrs rocmSupport {
PYTORCH_ROCM_ARCH = lib.concatStringsSep ";" torch.rocmArchs;
Expand Down
Loading