Skip to content

Refine build matrix and CUDA architecture specifications #172

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 5 commits into from
Sep 20, 2025
Merged

Refine build matrix and CUDA architecture specifications #172

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
dd73ce3
Updates architecture matrix to include only relevant architectures fo…
algo-home Sep 20, 2025
26bd08d
Updates default CUDA architectures to include only relevant versions …
algo-home Sep 20, 2025
735df58
Refines build matrix by removing ARM64 OS and updating Python version…
algo-home Sep 20, 2025
4d9cec0
Refines Python and Torch version specifications for improved compatib…
algo-home Sep 20, 2025
2d7f6b6
Updates Python and Torch version specifications for improved compatib…
algo-home Sep 20, 2025
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
6 changes: 3 additions & 3 deletions .github/workflows/publish.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -43,16 +43,16 @@ jobs:
matrix:
# Using ubuntu-22.04 instead of 24.04 for more compatibility (glibc). Ideally we'd use the
# manylinux docker image, but I haven't figured out how to install CUDA on manylinux.
os: [ubuntu-22.04, ubuntu-22.04-arm64]
python-version: ["3.9", "3.10", "3.11", "3.12", "3.13"]
os: [ubuntu-22.04]
python-version: ["3.9", "3.10", "3.11", "3.12"]
torch-version: ["2.5.1", "2.6.0", "2.7.1", "2.8.0"]
cuda-version: ["12.9.1"]
# We need separate wheels that either uses C++11 ABI (-D_GLIBCXX_USE_CXX11_ABI) or not.
# Pytorch wheels currently don't use it, but nvcr images have Pytorch compiled with C++11 ABI.
# Without this we get import error (undefined symbol: _ZN3c105ErrorC2ENS_14SourceLocationESs)
# when building without C++11 ABI and using it on nvcr images.
cxx11_abi: ["FALSE", "TRUE"]
arch: ["80", "86", "89", "90", "100", "120"]
arch: ["80", "90"]
include:
- torch-version: "2.9.0.dev20250904"
cuda-version: "13.0"
Expand Down
2 changes: 1 addition & 1 deletion setup.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,7 @@ def should_skip_cuda_build():

@functools.lru_cache(maxsize=None)
def cuda_archs():
return os.getenv("FLASH_DMATTN_CUDA_ARCHS", "80;86;89;90;100;120").split(";")
return os.getenv("FLASH_DMATTN_CUDA_ARCHS", "80;90").split(";")


def detect_preferred_sm_arch() -> Optional[str]:
Expand Down