Skip to content
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.

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

Initial support gfx950 #64

Merged
merged 12 commits into from
Jan 10, 2025
Merged

Initial support gfx950 #64

merged 12 commits into from
Jan 10, 2025

Conversation

xinyazhang
Copy link
Collaborator

@xinyazhang xinyazhang commented Jan 9, 2025

  • Supported under the name "Unidentified"
  • Bump the Triton compiler
  • Add pyaotriton.HipMemory to manage GPU memory when Torch is CPU-only
  • Add env var AOTRITON_TORCH_ONLY_USE_CPU to run unit tests with CPU-only Torch

Note: as initial support, there is no tuning database for gfx950

@xinyazhang xinyazhang marked this pull request as ready for review January 9, 2025 06:17
@xinyazhang
Copy link
Collaborator Author

xinyazhang commented Jan 9, 2025

All FOR_RELEASE=1 test/test_backward.py passed on real hardware.

@xinyazhang xinyazhang merged commit e8515a8 into main Jan 10, 2025
ethanwee1 pushed a commit that referenced this pull request Feb 26, 2025
* Supported under the name "Unidentified"
* Bump the Triton compiler
* Add `pyaotriton.HipMemory` to manage GPU memory when Torch is CPU-only
* Add env var `AOTRITON_TORCH_ONLY_USE_CPU` to run unit tests with CPU-only Torch

Note: as initial support, there is no tuning database for gfx950
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants