[benchmark_inference.py] Specify tp_size to StaticCache
#2784
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
Signed-off-by: Masaki Kozuki <[email protected]>
Contributor
There was a problem hiding this comment.
Pull request overview
This PR enhances tensor parallel support in the inference benchmark by specifying the tp_size parameter when initializing StaticCache for transformers >= 4.55. The changes also fix tensor parallel plan patterns to be more specific and add sanity checks to verify proper sharding.
Key Changes
- Fixed tensor parallel plan patterns from
*.layers.*tomodel.layers.*for more precise module matching - Added
tp_sizeparameter toStaticCacheinitialization to properly handle sharded KV heads in tensor parallel configurations - Added DTensor verification assertions for attention projection weights to ensure proper sharding
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
shino16
approved these changes
Dec 3, 2025
Collaborator
shino16
left a comment
shino16
left a comment
There was a problem hiding this comment.
Thanks! I remember I faced the same issue at some older commits, and I was wondering how it could be reproduced.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
What does this PR do?
KV values seem to be intact even when tensor parallel is enabled, thus specify
tp_sizein StaticCache.