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.

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

Jump to bottom

[WIP] Add gfx1100 support to AMD pytorch build #2642

Draft
wants to merge 1 commit into
base: main
Choose a base branch
from
Draft

[WIP] Add gfx1100 support to AMD pytorch build #2642

wants to merge 1 commit into from

Conversation

cazlo
Copy link

@cazlo cazlo commented Oct 13, 2024 •
edited
Loading

What does this PR do?

Fixes #2641

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you read the contributor guideline,
    Pull Request section?
  • Was this discussed/approved via a Github issue or the forum? Please add a link
    to it if that's the case.
  • Did you make sure to update the documentation with your changes? Here are the
    documentation guidelines, and
    here are tips on formatting docstrings.
  • Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

Reference

@mht-sharma
Copy link
Collaborator

Hi @cazlo, thanks for the PR. FYI, you would also need to modify the: vllm and flash-attention CK

I have not looked into the support for these GPUs in composable kernel and VLLM yet. But let me know if you face any issues.

@lhl
Copy link

lhl commented Dec 10, 2024

CK (and hence ROCm/flash-attention) does not support gfx1100 so it's best to just be disabled and set ROCM_USE_FLASH_ATTN_V2_TRITON=1? Note quite sure what vllm is used for, but to use Triton FA, you need to set VLLM_USE_TRITON_FLASH_ATTN=1. Also for PyTorch you may need TORCH_ROCM_AOTRITON_ENABLE_EXPERIMENTAL=1.

If you're looking to successfully build vLLM for ROCm, see: https://github.com/vllm-project/vllm/blob/main/Dockerfile.rocm

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.

Add AMD gfx110* support
3 participants
@cazlo @mht-sharma @lhl