[Klaud Cold] Update kimik2.5-int4-mi325x-vllm vLLM ROCm image to v0.24.0 / 将 kimik2.5-int4-mi325x-vllm 的 vLLM ROCm 镜像 升级至 v0.24.0

[functionstackx]

Summary

Update vLLM ROCm image from vllm/vllm-openai-rocm:v0.21.0 to vllm/vllm-openai-rocm:v0.24.0

Recipes touched: kimik2.5-int4-mi325x-vllm

中文说明

将 vLLM ROCm 镜像 从 vllm/vllm-openai-rocm:v0.21.0 升级至 vllm/vllm-openai-rocm:v0.24.0。涉及配置：kimik2.5-int4-mi325x-vllm。

Test plan

• full-sweep-fail-fast sweep passes.

🤖 Generated with Claude Code

[github-actions]

Thanks for the contribution! Please reach out to respective companies' CODEOWNER to fill in the latest PR_REVIEW_CHECKLIST.md before pinging core maintainer on Slack for review. In order for the signoff PR check bot to trigger, you must follow the PR_REVIEW_CHECKLIST.md template correctly, including the phrase As a PR reviewer and CODEOWNER, I have reviewed this and have.

For PR verification, add the full-sweep-enabled or full-sweep-fail-fast label to this PR — the benchmark sweep only runs on labeled PRs.

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. See GitHub's docs on re-running failed jobs

---

感谢你的贡献！请联系相应公司的 CODEOWNER 填写最新的 PR_REVIEW_CHECKLIST.md，然后再在 Slack 上联系核心维护者进行审阅。为了触发 signoff PR 检查机器人，你必须正确遵循 PR_REVIEW_CHECKLIST.md 模板，包括保留英文语句 As a PR reviewer and CODEOWNER, I have reviewed this and have。

如需进行 PR 验证，请为此 PR 添加 full-sweep-enabled 或 full-sweep-fail-fast 标签 — 基准测试 sweep 仅在带有标签的 PR 上运行。

PR 作者有责任确保合并后所有 GitHub Action 任务完全通过。 很多时候失败只是偶发抖动（flake），重新运行失败的任务即可解决。参见 GitHub 关于重新运行失败任务的文档

[github-actions]

Thanks for the contribution! Please reach out to respective companies' CODEOWNER to fill in the latest PR_REVIEW_CHECKLIST.md before pinging core maintainer on Slack for review. In order for the signoff PR check bot to trigger, you must follow the PR_REVIEW_CHECKLIST.md template correctly, including the phrase As a PR reviewer and CODEOWNER, I have reviewed this and have.

For PR verification, add the full-sweep-enabled or full-sweep-fail-fast label to this PR — the benchmark sweep only runs on labeled PRs.

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. See GitHub's docs on re-running failed jobs

---

感谢你的贡献！请联系相应公司的 CODEOWNER 填写最新的 PR_REVIEW_CHECKLIST.md，然后再在 Slack 上联系核心维护者进行审阅。为了触发 signoff PR 检查机器人，你必须正确遵循 PR_REVIEW_CHECKLIST.md 模板，包括保留英文语句 As a PR reviewer and CODEOWNER, I have reviewed this and have。

如需进行 PR 验证，请为此 PR 添加 full-sweep-enabled 或 full-sweep-fail-fast 标签 — 基准测试 sweep 仅在带有标签的 PR 上运行。

PR 作者有责任确保合并后所有 GitHub Action 任务完全通过。 很多时候失败只是偶发抖动（flake），重新运行失败的任务即可解决。参见 GitHub 关于重新运行失败任务的文档

[claude]

LGTM — mechanical vLLM ROCm image bump for a single recipe with matching perf-changelog entry; full-sweep-fail-fast will validate.

Extended reasoning...

Overview

This PR bumps the vLLM ROCm image for the kimik2.5-int4-mi325x-vllm recipe from vllm/vllm-openai-rocm:v0.21.0 to vllm/vllm-openai-rocm:v0.24.0 in configs/amd-master.yaml, and adds a corresponding entry in perf-changelog.yaml linking the PR. Two lines of substantive change total, both mechanical and following the well-established [Klaud Cold] image-bump pattern used repeatedly across this repo.

Security risks

None. This is a benchmark-config-only change: no runtime code, no auth/crypto/permissions surface, no CI logic. The image reference points to the same publicly published vllm/vllm-openai-rocm namespace already used by sibling recipes in the file.

Level of scrutiny

Very low. The change is a pure config tweak in a benchmark recipe scoped to one model+precision+SKU combination. The full-sweep-fail-fast label is applied, so the sweep will independently validate that the new image runs the recipe end-to-end — if v0.24.0 breaks the recipe, CI catches it before merge.

Other factors

The bug hunting system found no issues. Changelog entry format matches the surrounding entries. No CODEOWNER-gated code paths touched. No outstanding reviewer comments to address (only the standard bot reminder).