[Klaud Cold] [AMD] gpt-oss-fp4-mi355x (vllm): W4A8 moe optimizations and vllm image bump / gpt-oss-fp4-mi355x（vLLM）：W4A8 MoE 优化与 vLLM 镜像升级

[xiaohuguo2023]

• Image & model

  • Bump image to vllm/vllm-openai-rocm:nightly-68ee8300a047db78fb52bac477daaaac7be11216 (vLLM 0.23.1rc1, aiter 0.1.16.post2, triton 3.6.0).
  • Run amd/gpt-oss-120b-w-mxfp4-a-fp8 (W4A8), exercising the activation-quantized a8w4 Triton MoE path.

• a8w4 optimizations picked up via the image

  • aiter: updated A8W4 MoE GEMM tuning configs for better pipelining + an output-allocation correctness fix.
  • vLLM: hybrid CDNA4 MX-scale swizzle gate (swizzle ON for TP≤2, OFF for TP4/8).

• Extend the concs sweep coverage

[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]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28723312374
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28723312374

[functionstackx]

https://github.com/SemiAnalysisAI/InferenceX/actions/runs/28723312374/job/85184218900

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28755420485
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28755420485

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28759202756
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28759202756

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28759202756
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28759202756

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28790327187
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28790327187

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28790327187
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28790327187

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28790327187
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28790327187