[CI] run tests/compile/test_config.py

[zou3519]

Purpose

This test wasn't being run in CI

Test Plan

Run test locally.
Wait for CI, read logs

Test Result

Test passed locally

[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]

Code Review

This pull request correctly adds a missing test, tests/compile/test_config.py, to the CI pipeline. My review focuses on improving the maintainability of this CI step. I've suggested replacing the explicit list of test files with a single command to run all tests in the directory. This will prevent similar omissions in the future and make the CI configuration more robust.

[ProExpertProg]

Let's merge this as-is and we can organize better later

[yewentao256]

Looks good to me, thanks for the work!

[yewentao256]

LGTM, thanks for the work!

[mergify]

This pull request has merge conflicts that must be resolved before it can be
merged. Please rebase the PR, @zou3519.

https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/working-with-forks/syncing-a-fork

[ZJY0516]

This failed again due to

[2025-09-23T15:25:35Z] FAILED compile/test_silu_mul_quant_fusion.py::test_fusion_silu_and_mul_quant[True-TestSiluMulFp8QuantModel-dtype0-128-32] - AssertionError: Tensor-likes are not close!
--
  | [2025-09-23T15:25:35Z]
  | [2025-09-23T15:25:35Z] Mismatched elements: 65 / 128 (50.8%)
  | [2025-09-23T15:25:35Z] Greatest absolute difference: 0.0625 at index (0,) (up to 0.001 allowed)
  | [2025-09-23T15:25:35Z] Greatest relative difference: 0.007049560546875 at index (0,) (up to 0.001 allowed)

[ProExpertProg]

Yea not sure why it fails here but not on main or locally... Perhaps we just update the tolerance. Also the rmsnorm fusion test succeeds locally for me

[ZJY0516]

I don't believe adjusting the tolerance is a good idea. While it might hide the numerical issues in the short term, it's not a sustainable solution as it avoids addressing the root cause.

[ProExpertProg]

@ZJY0516 these tolerances are pretty arbitrary, we don't have a good reason for it to be what it currently is. That's why we rely on E2E lm_eval testing