[Klaud Cold] Update gptoss-fp4-b200-vllm vLLM image to v0.22.0

[functionstackx]

Summary

Update vLLM image from v0.21.0 to v0.22.0

Recipes touched: gptoss-fp4-b200-vllm

Test plan

• full-sweep-enabled sweep passes.

🤖 Generated with Claude Code

---

Note

Low Risk
Config-only image version bump for an existing benchmark recipe; no application logic or sweep parameters changed.

Overview
Bumps the gptoss-fp4-b200-vllm benchmark recipe to vllm/vllm-openai:v0.22.0 (from v0.21.0) in nvidia-master.yaml, and records the change in perf-changelog.yaml with config key gptoss-fp4-b200-vllm and PR link.

No model, runner, precision, or sweep search-space changes—only the container image version used for this GPT-OSS FP4 on B200 vLLM config.

^{Reviewed by Cursor Bugbot for commit 6f1e44b. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

• https://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

• https://github.com/vllm-project/recipes
• https://github.com/sgl-project/sglang/tree/main/docs_new

PR authors are responsible for ensuring that after merging, all GitHub Action jobs fully pass. A lot of the time, failures are just flakes and simply re-running the failed jobs will fix it. If re-running failed jobs is attempted, PR authors are responsible for ensuring it passes. See GitHub's docs on re-running failed jobs: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/re-run-workflows-and-jobs#re-running-failed-jobs-in-a-workflow

As a rule of thumb, generally, PR authors should request a review & get a PR approval from the respective companies' CODEOWNERS before requesting a review from core maintainers.

If additional help is needed, PR authors can reach out to core maintainers over Slack.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26696224818
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26696224818

[claude]

LGTM — straightforward vLLM image version bump (v0.21.0 → v0.22.0) for gptoss-fp4-b200-vllm with matching perf-changelog entry.

Extended reasoning...

Overview

This PR updates the vLLM container image tag from v0.21.0 to v0.22.0 for the gptoss-fp4-b200-vllm recipe in .github/configs/nvidia-master.yaml, plus a corresponding entry appended to perf-changelog.yaml. Only two lines of functional change (the image tag) plus a documentation/changelog block.

Security risks

None. This is a config-only change that bumps a pinned image tag used by CI sweep jobs. No code paths, auth, crypto, or permissions are touched.

Level of scrutiny

Low. Image version bumps are mechanical and follow an established repo pattern (see recent commits like 1786ed4 which is this exact PR pattern, and 48c1840 for an SGLang image bump). Correctness of the image itself is validated by the full-sweep-enabled label which triggers a sweep — a human review wouldn't add signal beyond what the sweep run provides.

Other factors

The bug hunting system found no issues. The PR follows the same shape as other recently merged image-bump PRs in the repo. The changelog entry is appropriately scoped (single config key, single-line description). The recipe-reminder bot comment is the standard auto-posted reminder, not a substantive review concern.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26696226421
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26696226421

[functionstackx]

/reuse-sweep-run

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26699237645
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26699237645