Fixing Chunked Prefill Test.

[Alexei-V-Ivanov-AMD]

Fixing Chunked Prefill Test.

Signed-off-by: Alexei V. Ivanov alexei.ivanov@amd.com

[github-actions]

👋 Hi! Thank you for contributing to the vLLM project.

💬 Join our developer Slack at https://slack.vllm.ai to discuss your PR in #pr-reviews, coordinate on features in #feat- channels, or join special interest groups in #sig- channels.

Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can either: Add ready label to the PR or enable auto-merge.

🚀

[gemini-code-assist]

Summary of Changes

Hello @Alexei-V-Ivanov-AMD, 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 aims to fix failures in the Chunked Prefill Test suite by introducing conditional skips for tests that rely on features or attention backends not supported on ROCm platforms. This ensures the test suite's reliability and correctness across different hardware environments by avoiding unsupported test configurations.

Highlights

• Test Stability on ROCm: I've added pytest.mark.skipif decorators to several test functions within test_chunked_prefill.py to prevent them from running on ROCm platforms when specific hardware-dependent features or attention backends are not supported. This ensures the test suite passes correctly on ROCm without attempting unsupported configurations.
• FLASHINFER Backend Compatibility: The FLASHINFER attention backend is now conditionally skipped for both test_models (lines 53-56) and test_models_distributed (lines 109-112) when the test environment is detected as ROCm, as FLASHINFER is not supported on this platform.
• FP8 KV Cache Compatibility: The test_models_with_fp8_kv_cache test (lines 187-188) is now skipped on ROCm platforms because the underlying machete_prepack_B functionality required for FP8 KV cache is not supported there.

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 in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

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 issue 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 is currently in preview and 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 to provide feedback.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

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 effectively addresses test failures on ROCm platforms by conditionally skipping test configurations that rely on unsupported features. The changes are clear and use standard pytest mechanisms.