[BugFix] fix dsv4 piecewise scenario

[zouyida2052]

What this PR does / why we need it?

In PR#43746,while VLLM_USE_BREAKABLE_CUDAGRAPH is detacted as True, compilation_config.mode would be None even if we set it, thus server runs in eager mode.

Does this PR introduce any user-facing change?

no

How was this patch tested?

• vLLM version: v0.20.2
• vLLM main: vllm-project/vllm@9090368

[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 refactors the configuration logic for Ascend platforms to ensure that the breakable cudagraph feature is correctly disabled. By moving the environment variable override to the module initialization, the change prevents potential conflicts introduced by recent upstream changes in vLLM regarding model compilation support.

Highlights

• Configuration Adjustment: Moved the disabling of VLLM_USE_BREAKABLE_CUDAGRAPH from the platform defaults method to the module initialization phase in vllm_ascend/platform.py.
• Compatibility Fix: Ensures that the breakable cudagraph feature remains disabled on Ascend platforms to maintain compatibility with DeepSeek V4 model requirements.

New Features

🧠 You can now enable Memory (public preview) 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 the 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 counterproductive. 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. ↩

[gemini-code-assist]

Code Review

This pull request moves the logic that disables breakable CUDA graphs on Ascend from the apply_config_platform_defaults method to the module level of vllm_ascend/platform.py. The reviewer suggests also setting the environment variable os.environ["VLLM_USE_BREAKABLE_CUDAGRAPH"] = "False" to ensure that direct environment variable lookups respect this override, and provides a suggested PR title and summary in accordance with the repository's style guide.

[gemini-code-assist]

To ensure that any direct environment variable lookups (e.g., via os.getenv or os.environ.get) also respect this override, and to handle cases where other modules might import or check the environment variable directly, we should set both os.environ["VLLM_USE_BREAKABLE_CUDAGRAPH"] = "False" and envs_vllm.VLLM_USE_BREAKABLE_CUDAGRAPH = False.

---

Suggested PR Title:

[Ops][BugFix] fix dsv4 piecewise scenario

Suggested PR Summary:

### What this PR does / why we need it?

This PR force disables the breakable CUDA graph PIECEWISE path on Ascend at the module level of `vllm_ascend/platform.py`. This is necessary because DeepSeek V4 model classes no longer carry `@support_torch_compile`, causing vLLM to auto-enable the breakable CUDA graph PIECEWISE path, which is not yet supported on Ascend.

### Does this PR introduce _any_ user-facing change?

No.

### How was this patch tested?

CI/CD testing.

os.environ["VLLM_USE_BREAKABLE_CUDAGRAPH"] = "False"
envs_vllm.VLLM_USE_BREAKABLE_CUDAGRAPH = False

References

1. Pull Request Summary Style Guide: Output the generated title and summary in markdown code blocks. ^(link)

[github-actions]

👋 Hi! Thank you for contributing to the vLLM Ascend project. The following points will speed up your PR merge:‌‌

• A PR should do only one thing, smaller PRs enable faster reviews.
• Every PR should include unit tests and end-to-end tests ‌to ensure it works and is not broken by other future PRs.
• Write the commit message by fulfilling the PR description to help reviewer and future developers understand.

If CI fails, you can run linting and testing checks locally according Contributing and Testing.