[FA] Unify Base + Opt FWD kernels #233
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
codingwithsurya
had a problem deploying
to
docker-s3-upload
GitHub Actions
Error
codingwithsurya
had a problem deploying
to
docker-s3-upload
GitHub Actions
Error
|
This pull request was exported from Phabricator. Differential Revision: D75388323 |
Summary: This PR consolidates redundant TMA attention kernels into a unified implementation. Previously, `_attn_fwd_tma` and `_attn_fwd_tma_ws` contained duplicate code (mainly the TMA descriptors) and didn't leverage the existing `ENABLE_WS` flag. I've merged the redundant kernels into a single `_attn_fwd_tma_unified` kernel. We now use the `ENABLE_WS` flag to toggle between regular and warp-specialized execution. Changes: * Merged both kernels into `_attn_fwd_tma_unified` kernel with handling of regular and warp-specialized paths * Utilized existing `ENABLE_WS` parameter to control warp specialization * Unified TMA descriptor creation logic Differential Revision: D75307125
Summary: Separated the TMA kernel variant handling into distinct code paths rather than using a conditional parameter. Changed from a unified approach with a dynamic `is_warp_specialized` flag to explicit separate conditions for `tma` and `tma_ws` variants. This improves code clarity by making the execution path more explicit + makes it easier for compiler to optimize. Differential Revision: D75308966
Summary: I merged the "base" (`_attn_fwd`) and "opt" (`_attn_fwd_opt`) variants of the forward pass of the attention kernel into a single `_attn_fwd_unified` kernel to streamline the codebase and avoid redundancy. Both the base and opt kernels were great candidates to be merged since they were nearly identical code paths. I still need to integrate the warp spec path (`attn_fwd_ws`) after some bug-fixing. For now it's still available as its own autotuned kernel. Differential Revision: D75388323
facebook-github-bot
force-pushed
the
export-D75388323
branch
from
0c93d76 to
5566672
Compare
facebook-github-bot
had a problem deploying
to
docker-s3-upload
GitHub Actions
Error
facebook-github-bot
had a problem deploying
to
docker-s3-upload
GitHub Actions
Error
|
This pull request was exported from Phabricator. Differential Revision: D75388323 |
codingwithsurya
changed the title
codingwithsurya
changed the title
codingwithsurya
changed the title
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:
I merged the "base" (
_attn_fwd) and "opt" (_attn_fwd_opt) variants of the forward pass of the attention kernel into a single_attn_fwd_unifiedkernel to streamline the codebase and avoid redundancy. Both the base and opt kernels were great candidates to be merged since they were nearly identical code paths.This PR also includes diffs from #232 due to Github Export from Phabricator. #232 should be merged before this.
Test Plan:
Unit Tests and Benchmarking
Differential Revision: D75388323