Enable rowwise scaling for DeepGemm #3874
Closed
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
Summary: This diff adds [ngimel's support for DeepGemm rowwise scaling](https://github.com/ngimel/DeepGEMM/tree/rowwise) to our fbcode copy. It also includes a few deepgemm updates that allow operation on M<128, which is important for any real use case. Performance is increased considerably by the use of rowwise scaling, especially in memory bound cases. Notably, this makes DeepGemm the premier solution for slow accumulation as it now overall outperforms cublas + unfused rowwise scaling. {F1976375307} Differential Revision: D71748927
|
This pull request was exported from Phabricator. Differential Revision: D71748927 |
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify site configuration. |
|
This pull request has been merged in ee4c88b. |
q10
pushed a commit
to q10/FBGEMM
that referenced
this pull request
Apr 10, 2025
Summary: Pull Request resolved: facebookresearch/FBGEMM#964 X-link: pytorch#3874 This diff adds [ngimel's support for DeepGemm rowwise scaling](https://github.com/ngimel/DeepGEMM/tree/rowwise) to our fbcode copy. It also includes a few deepgemm updates that allow operation on M<128, which is important for any real use case. Performance is increased considerably by the use of rowwise scaling, especially in memory bound cases. Notably, this makes DeepGemm the premier solution for slow accumulation as it now overall outperforms cublas + unfused rowwise scaling. {F1976375307} Reviewed By: jianyuh Differential Revision: D71748927 fbshipit-source-id: 87e287a2cec284bd8fd7c5e80603065a0d662f53
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.
Summary:
This diff adds ngimel's support for DeepGemm rowwise scaling to our fbcode copy. It also includes a few deepgemm updates that allow operation on M<128, which is important for any real use case. Performance is increased considerably by the use of rowwise scaling, especially in memory bound cases. Notably, this makes DeepGemm the premier solution for slow accumulation as it now overall outperforms cublas + unfused rowwise scaling.
{F1976375307}
Differential Revision: D71748927