perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-disagg harness

[cquil11]

Summary

• Add a perf changelog entry for dsv4-fp4-gb300-dynamo-sglang-agentic-disagg.
• Refresh the submission with the up-to-date AgentX harness.

Validation

• Parsed perf-changelog.yaml successfully.
• git diff --check passes.

[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

[claude]

Claude finished @cquil11's task in 1m 19s —— View job

---

PR Review

• Gather context (diff vs origin/main)
• Verify perf-changelog entry position and format
• Post review

LGTM - no blocking issues found

Verified:

• The new entry is appended at the end of perf-changelog.yaml (after the previous last entry for PR perf(agentx): refresh Qwen3.5 GB300 recipes with correct metrics #2477), preserving chronological order. The whitespace-only line at EOF is pre-existing on main, not introduced here.
• Entry structure (config-keys, scenario-type, description, pr-link) matches the schema used by existing entries, and git diff --check passes.
• The config key dsv4-fp4-gb300-dynamo-sglang-agentic-disagg exists in configs/nvidia-master.yaml (line 8039).
• This is a changelog-only PR (no master-config changes), so no cross-file update is required.

[claude]

LGTM — trivial changelog-only entry, no config or script changes.

Extended reasoning...

Overview

This PR adds a single new entry to perf-changelog.yaml documenting a refresh of the dsv4-fp4-gb300-dynamo-sglang-agentic-disagg AgentX harness. No benchmark configs, scripts, or code are touched — it is a pure metadata/documentation addition following the exact same config-keys/scenario-type/description/pr-link structure as every other entry in the file.

Security risks

None. The file is a YAML changelog with no executable content, secrets, or credentials.

Level of scrutiny

Minimal scrutiny is appropriate — this is a mechanical, additive-only change to a log file, not a change to any harness logic, launcher routing, or master config.

Other factors

The bug hunting system found no issues. Two candidate concerns were raised and ruled out: use of the full-sweep-enabled label instead of the recommended full-sweep-fail-fast, and the changelog referencing a harness refresh with no accompanying config/script diff — neither is a functional bug given this PR's scope is changelog-only. The new entry's YAML structure and indentation match the surrounding entries exactly.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31150507931
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=31150507931

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31150507931
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=31150507931

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31152168360
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=31152168360

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31153421537
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=31153421537

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31153886144
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=31153886144

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31155359606
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=31155359606