Enable AOTriton BWD V3 API #382
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
Micky774
requested review from
ipanfilo,
wangye805 and
wenchenvincent
as code owners
ipanfilo
reviewed
Nov 26, 2025
transformer_engine/common/fused_attn_rocm/fused_attn_aotriton.cpp
Outdated
Show resolved
Hide resolved
ipanfilo
approved these changes
Dec 1, 2025
Contributor
Author
|
Note this PR is not merge ready -- there are still some CI failures from the level 3 run that need to be corrected. |
Micky774
force-pushed
the
zain/aotriton-bwd
branch
from
ff336b5 to
51da203
Compare
Contributor
Author
|
CI failures have been addressed (thanks @xinyazhang for your help!). Ready for review @wenchenvincent @ipanfilo @wangye805 |
ipanfilo
requested changes
Feb 11, 2026
| RoPE, | ||
| is_training, | ||
| ) | ||
| if len(fused_attn_backends) == 1: |
ipanfilo
approved these changes
Feb 11, 2026
| is_training, | ||
| ) | ||
| elif len(fused_attn_backends) == 2: | ||
| elif len(fused_attn_backends) == 2 and IS_HIP_EXTENSION: |
Collaborator
There was a problem hiding this comment.
IS_HIP_EXTENSION here is redundant
| set(__AOTRITON_SUFFIX "_TEprivate") | ||
|
|
||
| if(NOT DEFINED AOTRITON_PATH) | ||
| # If AOTRITON_PATH is not provided, we proceed to build the runtime |
Collaborator
There was a problem hiding this comment.
Your original changes used env variable to control it. On the other hand this feature seems unused
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.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
This is a follow-up to the earlier AOTriton update PR (#360). This one focuses on adopting the AOTriton V3 API for the BWD pass, including introducing a
LazyTensorwrapper implementation (driven by an eager tensor as usual).Follow-up for #360
Type of change
Changes
Please list the changes introduced in this PR:
Checklist: