Enable the qwen3 test

[Shunkangz]

Motivation

Modifications

Accuracy Tests

Benchmarking and Profiling

Checklist

• Format your code according to the Format code with pre-commit.
• Add unit tests according to the Run and add unit tests.
• Update documentation according to Write documentations.
• Provide accuracy and speed benchmark results according to Test the accuracy and Benchmark the speed.
• Follow the SGLang code style guidance.

Review Process

1. Ping Merge Oncalls to start the PR flow. 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.
   • /tag-run-ci-label, /rerun-failed-ci, /tag-and-rerun-ci
4. After green CI and required approvals, ask Merge Oncalls to merge.

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request focuses on enabling the Qwen3 30B model test within the CI pipeline and refining the Qwen3 Mixture of Experts (MoE) model's forward pass. The changes ensure that expert parallel all-reduce operations are correctly applied under specific conditions, which is crucial for the model's distributed execution and accuracy.

Highlights

• Qwen3 MoE Model Update: Introduced a conditional moe_expert_parallel_all_reduce operation within the forward_normal function of the Qwen3 MoE model, specifically when expert parallelism is enabled and all-reduce fusion is not active. This ensures proper handling of expert outputs in distributed environments.
• Qwen3 Test Re-enablement: The test_qwen3_30b.py test case, which was previously disabled, has been re-enabled by removing the disabled flag from its register_cuda_ci decorator. This allows the Qwen3 30B model test to run as part of the CI pipeline.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[Shunkangz]

/tag-run-ci-label

[gemini-code-assist]

Code Review

This pull request enables a test for qwen3 and includes a fix for expert parallelism in the MoE layer. The fix adds an all-reduce operation for expert parallelism. However, the condition under which this all-reduce is performed seems incorrect, as it is tied to the fusion of the tensor parallelism all-reduce, which can lead to incorrect results when fusion is enabled. I've suggested a correction to ensure the expert parallelism all-reduce is always performed when needed.

[Fridge003]

/tag-and-rerun-ci

[Fridge003]

/rerun-stage stage-c-test-4-gpu-h100

[github-actions]

✅ Triggered stage-c-test-4-gpu-h100 to run independently (skipping dependencies).

[github-actions]

🔗 View workflow run

[Fridge003]

H100 tests passed https://github.com/sgl-project/sglang/actions/runs/23463165796/job/68305906001