expand to arbitrary K size for MXFP4#10
Open
moizyousufi wants to merge 3 commits intoIST-DASLab:mainfrom
Open
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
The K size for MXFP4 was limited to 128. I noticed that natively we can expand to 256.
As for expanding >256, I have created a v2 API such that ≤256 defaults to legacy and >256 defaults to v2.
The v2 API
gemm_op.get_workspace_size. Now this will let QuTLASS scale up as high as you need.WarpShape.N == Kso there isn't accumulation across K-tiles and the scale factors are computed once per warpOther Notes
Also, I added a conda env for this so that there can be better/more consistent testing on this.