Adding support for top_k=8 to index_shuffling moe kernel #5142
+181
−35
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
Contributor
|
@metastableB has exported this pull request. If you are a Meta employee, you can view the originating Diff in D86728049. |
metastableB
added a commit
to metastableB/FBGEMM-1
that referenced
this pull request
Nov 17, 2025
Summary: X-link: facebookresearch/FBGEMM#2142 - Adds support for top_k=8 argument to the `index_shuffling` kernel, used as part of token shuffling moe implementations. Modifies tiling/dispatch logic to enable this. - Adds unit test for top_k=8. Differential Revision: D86728049
metastableB
force-pushed
the
export-D86728049
branch
from
e0b769f to
ee1b408
Compare
Summary:
- Existing `index_shuffling` kernel in fbgemm supports `top_k` values of `{1, 2, 4}`. This diff adds support for top_k=8.
- Refactors tiling/dispatch logic to enable the `top_k=8` case
- Adds unit test for k=8 and adds another additional test to explicitly test the tile size boundaries.
Reviewed By: jasonjk-park
Differential Revision: D86728049
metastableB
force-pushed
the
export-D86728049
branch
from
ee1b408 to
65413cc
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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:
X-link: https://github.com/facebookresearch/FBGEMM/pull/2142
Adds support for top_k=8 argument to the
index_shufflingkernel, used as part of token shuffling moe implementations. Modifies tiling/dispatch logic to enable this.Adds unit test for top_k=8.
Differential Revision: D86728049