Skip to content

[ROCm][BugFix]Fix get_cu_count in rocm_aiter_fa.py #28618

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
merged 2 commits into from
Nov 13, 2025
Merged

[ROCm][BugFix]Fix get_cu_count in rocm_aiter_fa.py #28618

merged 2 commits into from
Nov 13, 2025

Conversation

@ganyi1996ppo
Copy link
Contributor

@ganyi1996ppo ganyi1996ppo commented Nov 13, 2025 •
edited by github-actions bot
Loading

Purpose

Sorry for missing this PR #28383 when fix get_cu_count issue, please take a look again
@tjtanaa @DarkLight1337

Test Plan

Test Result

Essential Elements of an Effective PR Description Checklist
  • The purpose of the PR, such as "Fix some issue (link existing issues this PR will resolve)".
  • The test plan, such as providing test command.
  • The test results, such as pasting the results comparison before and after, or e2e results
  • (Optional) The necessary documentation update, such as updating supported_models.md and examples for a new model.
  • (Optional) Release notes update. If your change is user facing, please update the release notes draft in the Google Doc.

@ganyi1996ppo
fix get_cu_count in rocm_aiter_fa.py
a26d5ec 
Signed-off-by: ganyi <ygan@amd.com>
@mergify mergify bot added rocm Related to AMD ROCm v1 labels Nov 13, 2025
gemini-code-assist[bot]
gemini-code-assist bot reviewed Nov 13, 2025
View reviewed changes
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request fixes a bug in rocm_aiter_fa.py where an incorrect method was used to get the number of compute units on ROCm platforms. The change to use get_cu_count is correct. However, I've identified a performance issue where this function would be called repeatedly in a loop. I've provided a suggestion to cache the result to improve performance.

@DarkLight1337 DarkLight1337 enabled auto-merge (squash) November 13, 2025 06:44
@github-actions github-actions bot added the ready ONLY add when PR is ready to merge/full CI is needed label Nov 13, 2025
@tjtanaa tjtanaa mentioned this pull request Nov 13, 2025
Closed
5 tasks
@DarkLight1337 DarkLight1337 merged commit 8da2f28 into vllm-project:main Nov 13, 2025
49 checks passed
geodavic pushed a commit to geodavic/vllm that referenced this pull request Nov 16, 2025
@ganyi1996ppo @geodavic
[ROCm][BugFix]Fix get_cu_count in rocm_aiter_fa.py (vllm-project#28618
0ec783a 
)

Signed-off-by: ganyi <ygan@amd.com>
Signed-off-by: George D. Torres <gdavtor@gmail.com>
bwasti pushed a commit to bwasti/vllm that referenced this pull request Nov 17, 2025
@ganyi1996ppo @bwasti
[ROCm][BugFix]Fix get_cu_count in rocm_aiter_fa.py (vllm-project#28618
335adcc 
)

Signed-off-by: ganyi <ygan@amd.com>
Signed-off-by: Bram Wasti <bwasti@meta.com>
devpatelio pushed a commit to SumanthRH/vllm that referenced this pull request Nov 29, 2025
@ganyi1996ppo @devpatelio
[ROCm][BugFix]Fix get_cu_count in rocm_aiter_fa.py (vllm-project#28618
1422e28 
)

Signed-off-by: ganyi <ygan@amd.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@DarkLight1337 DarkLight1337 DarkLight1337 approved these changes

@tjtanaa tjtanaa Awaiting requested review from tjtanaa tjtanaa is a code owner

+1 more reviewer

@gemini-code-assist gemini-code-assist[bot] gemini-code-assist[bot] left review comments

Reviewers whose approvals may not affect merge requirements

Assignees

No one assigned

Labels

ready ONLY add when PR is ready to merge/full CI is needed rocm Related to AMD ROCm v1

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@ganyi1996ppo @DarkLight1337