[CUTLASS] Use custom copy of cutlass to enable FP8 Grouped Gemm. #3649
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
|
@jwfromm has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
jiawenliu64
approved these changes
Jan 31, 2025
q10
approved these changes
Jan 31, 2025
avbokovoy
pushed a commit
to ROCm/FBGEMM
that referenced
this pull request
Feb 14, 2025
Summary: X-link: https://github.com/facebookresearch/FBGEMM/pull/726 To support MOE models, we need to enable FP8 rowwise grouped gemm in FBGEMM. One missing piece to do this is support for rowwise scaling in cutlass. We have enabled this feature in a custom copy of cutlass but getting it into mainline will take a while. As a temporary measure, we can point FBGEMM to a custom copy of cutlass instead. Once the feature lands in mainline and we bump our support, we can go back to using the main repo. Pull Request resolved: pytorch#3649 Reviewed By: q10, jiawenliu64 Differential Revision: D68967944 Pulled By: jwfromm fbshipit-source-id: 3e4625227ba6c33cf0478811fc9a8d40af361612
q10
pushed a commit
to q10/FBGEMM
that referenced
this pull request
Apr 10, 2025
Summary: Pull Request resolved: facebookresearch/FBGEMM#726 To support MOE models, we need to enable FP8 rowwise grouped gemm in FBGEMM. One missing piece to do this is support for rowwise scaling in cutlass. We have enabled this feature in a custom copy of cutlass but getting it into mainline will take a while. As a temporary measure, we can point FBGEMM to a custom copy of cutlass instead. Once the feature lands in mainline and we bump our support, we can go back to using the main repo. X-link: pytorch#3649 Reviewed By: q10, jiawenliu64 Differential Revision: D68967944 Pulled By: jwfromm fbshipit-source-id: 3e4625227ba6c33cf0478811fc9a8d40af361612
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
To support MOE models, we need to enable FP8 rowwise grouped gemm in FBGEMM. One missing piece to do this is support for rowwise scaling in cutlass. We have enabled this feature in a custom copy of cutlass but getting it into mainline will take a while. As a temporary measure, we can point FBGEMM to a custom copy of cutlass instead. Once the feature lands in mainline and we bump our support, we can go back to using the main repo.