MXFP4 Workgroup Reordering Based on SP3 Kernel#914
Open
suryajasper wants to merge 4 commits intoiree-org:mainfrom
Open
MXFP4 Workgroup Reordering Based on SP3 Kernel#914suryajasper wants to merge 4 commits intoiree-org:mainfrom
suryajasper wants to merge 4 commits intoiree-org:mainfrom
Conversation
Signed-off-by: Surya Jasper <45545431+suryajasper@users.noreply.github.com>
Signed-off-by: Surya Jasper <45545431+suryajasper@users.noreply.github.com>
Signed-off-by: Surya Jasper <45545431+suryajasper@users.noreply.github.com>
Signed-off-by: Surya Jasper <45545431+suryajasper@users.noreply.github.com>
f3953d5 to
7b5735b
Compare
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
MXFP4 Workgroup Reordering
Remaps the workgroup grid by creating groups of 32 tiles along the N dimension. This way, the first 32 workgroups in launch order are assigned along the N dimension on the same M row (0), then the next 32 to the next M row (1), and so on until wrapping around back to the first row.
Through this grouping, consecutive workgroups are able to reuse the same A row slices rather than spreading over a larger, sparse footprint. This improves L2 cache reuse since workgroups running close together in launch order work on the same row of output tiles.
The index remapping is calculated through two cases: