Skip to content

[qwen 3.8 next] reuse old cuda stream instead of endlessly creating streams - #39474

Merged
Qiaolin-Yu merged 3 commits into
mainfrom
qwen_stream
Sep 15, 2026
Merged

Qiaolin-Yu merged 3 commits into
mainfrom
qwen_stream

Conversation

@Qiaolin-Yu

@Qiaolin-Yu Qiaolin-Yu commented Sep 14, 2026 •

Copy link
Copy Markdown
Collaborator

Motivation

prev
image

after
image

Modifications

Accuracy Tests

Speed Tests and Profiling

Checklist

Review and Merge Process

  1. Ping Merge Oncalls to start the process. See the PR Merge Process.
  2. Get approvals from CODEOWNERS and other reviewers.
  3. Trigger CI tests with comments or contact authorized users to do so.
    • Common commands include /tag-and-rerun-ci, /tag-run-ci-label, /rerun-failed-ci
  4. After green CI and required approvals, ask Merge Oncalls or people with Write permission to merge the PR.

CI States

Latest PR Test (Base): ❌ Run #34938198301
Latest PR Test (Extra): ❌ Run #34938197926
Latest PR Test (AMD ROCm 10): ❌ Run #34938197976

@Qiaolin-Yu

Copy link
Copy Markdown
Collaborator Author

/rerun-test test/registered/e2e/models/test_qwen4_exp_models.py

@github-actions

github-actions Bot commented Sep 15, 2026 •

Copy link
Copy Markdown
Contributor

Results for /rerun-test test/registered/e2e/models/test_qwen4_exp_models.py:

🚀 4-gpu-b200 (1 test): ❌ View workflow run

cd test/ && python3 registered/e2e/models/test_qwen4_exp_models.py

@YAMY1234 YAMY1234 left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

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

LGTM

@Qiaolin-Yu

Copy link
Copy Markdown
Collaborator Author

/rerun-test test/registered/e2e/models/test_qwen4_exp_models.py

@github-actions

Copy link
Copy Markdown
Contributor

⚠️ CI Maintenance Mode is Active

The CI infrastructure is currently under maintenance. All PR CI runs are paused until maintenance is complete. Merging non-CI-fix PRs is prohibited during maintenance mode.

Follow issue #21065 for status updates. Re-run was not dispatched.

@Qiaolin-Yu

Copy link
Copy Markdown
Collaborator Author

/rerun-test test/registered/e2e/models/test_qwen4_exp_models.py

@github-actions

github-actions Bot commented Sep 15, 2026 •

Copy link
Copy Markdown
Contributor

Results for /rerun-test test/registered/e2e/models/test_qwen4_exp_models.py:

🚀 4-gpu-b200 (1 test): ✅ View workflow run

cd test/ && python3 registered/e2e/models/test_qwen4_exp_models.py

@Qiaolin-Yu
Qiaolin-Yu merged commit 2c0a709 into main Sep 15, 2026
107 of 123 checks passed
@Qiaolin-Yu
Qiaolin-Yu deleted the qwen_stream branch September 15, 2026 08:36
a-died-fish added a commit to endless-frontier/sglang that referenced this pull request Sep 17, 2026
Picks up the upstream QSA compress-gather fix (sgl-project#38346, relanded in sgl-project#39446) and
the qwen4_exp alt-stream reuse (sgl-project#39474). dev was missing both because it
contains the 2026-09-14 bulk revert (66c7bc8).

The endless-frontier deployment recipes are dev-only paths, so this merge is
conflict-free.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants