Add Paged Attention to FMHA Cutlass Blackwell Forward kernel for fixed length #4999
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 project configuration. |
|
@sarithad-meta has exported this pull request. If you are a Meta employee, you can view the originating Diff in D84023396. |
sarithad-meta
force-pushed
the
export-D84023396
branch
from
7ddeab4 to
d007740
Compare
sarithad-meta
added a commit
to sarithad-meta/FBGEMM-1
that referenced
this pull request
Oct 13, 2025
…d length (pytorch#4999) Summary: X-link: facebookresearch/FBGEMM#2013 Added paged attention support to FMHA FWD blackwell kernel. 1. Added support for fixed length case. 2. Added support for 2 cases: a) page_block_size = N tile size b) page_block_size > N 3. Added unit test, test_paged_forward. Next steps: 1. Test the performance of fixed length case. 2. Add support for variable length case to FWD kernel. 3. Add support for small page sizes to FWD kernel. 4. Add paged attention support for decode. Differential Revision: D84023396
sarithad-meta
force-pushed
the
export-D84023396
branch
from
d007740 to
3dde689
Compare
sarithad-meta
added a commit
to sarithad-meta/FBGEMM-1
that referenced
this pull request
Oct 16, 2025
…d length (pytorch#4999) Summary: X-link: facebookresearch/FBGEMM#2013 Added paged attention support to FMHA FWD blackwell kernel. 1. Added support for fixed length case. 2. Added support for 2 cases: a) page_block_size = N tile size b) page_block_size > N 3. Added unit test, test_paged_forward. Next steps: 1. Test the performance of fixed length case. 2. Add support for variable length case to FWD kernel. 3. Add support for small page sizes to FWD kernel. 4. Add paged attention support for decode. Reviewed By: Aya-ZIbra, sijiac Differential Revision: D84023396
sarithad-meta
force-pushed
the
export-D84023396
branch
from
3dde689 to
0f54e39
Compare
sarithad-meta
added a commit
to sarithad-meta/FBGEMM-1
that referenced
this pull request
Oct 17, 2025
…d length (pytorch#4999) Summary: Pull Request resolved: pytorch#4999 X-link: https://github.com/facebookresearch/FBGEMM/pull/2013 Added paged attention support to FMHA FWD blackwell kernel. 1. Added support for fixed length case. 2. Added support for 2 cases: a) page_block_size = N tile size b) page_block_size > N 3. Added unit test, test_paged_forward. Next steps: 1. Test the performance of fixed length case. 2. Add support for variable length case to FWD kernel. 3. Add support for small page sizes to FWD kernel. 4. Add paged attention support for decode. Reviewed By: Aya-ZIbra, sijiac Differential Revision: D84023396
…d length (pytorch#4999) Summary: Pull Request resolved: pytorch#4999 X-link: https://github.com/facebookresearch/FBGEMM/pull/2013 Added paged attention support to FMHA FWD blackwell kernel. 1. Added support for fixed length case. 2. Added support for 2 cases: a) page_block_size = N tile size b) page_block_size > N 3. Added unit test, test_paged_forward. Next steps: 1. Test the performance of fixed length case. 2. Add support for variable length case to FWD kernel. 3. Add support for small page sizes to FWD kernel. 4. Add paged attention support for decode. Reviewed By: Aya-ZIbra, sijiac Differential Revision: D84023396
sarithad-meta
force-pushed
the
export-D84023396
branch
from
0f54e39 to
298beb6
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/2013
Added paged attention support to FMHA FWD blackwell kernel.
Next steps:
Differential Revision: D84023396