[AMD][MI35X] 0706 DSV4 sglang

[1am9trash]

Successful run:
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28838628949
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28838628949

Change:

• "Bump image to lmsysorg/sglang-rocm:v0.5.14-rocm720-mi35x-20260706"
• "Clean the export envs"
• "Enable two batch overlap"

[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-fail-fast label (strongly recommended) to this PR — the benchmark sweep only runs on labeled PRs. Use full-sweep-enabled only if you need matrix jobs to keep running past a failure.

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-fail-fast 标签（强烈推荐）— 基准测试 sweep 仅在带有标签的 PR 上运行。仅当需要矩阵任务在失败后继续运行时才使用 full-sweep-enabled。

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=28784738508
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28784738508

[functionstackx]

@1am9trash can u update MTP please? it been 5 weeks since last mi355 dsv4 MTP update and within that 5 weeks, yall submitted like 6 different non-MTP updates but no MTP updates? MTP dsv4 is what is most relveant

https://github.com/SemiAnalysisAI/InferenceX/commits/6633e17c2e4b8eaad1db8995415d47f5436f9e3b/benchmarks/single_node/dsv4_fp4_mi355x_sglang_mtp.sh?browsing_rename_history=true&new_path=benchmarks/single_node/fixed_seq_len/dsv4_fp4_mi355x_sglang_mtp.sh&original_branch=08713fb7db07ff24489695b0aed05773b6fa4415

#1824 (review)

[1am9trash]

Hi, @functionstackx
I've opened another PR for SGLang V4 MTP.
PR: #2108

Just a heads-up, we still have a few PRs related to V4 MTP perf that haven't been merged yet, so this may need a follow-up update later.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=28838628949
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=28838628949

[1am9trash]

/reuse-sweep-run