perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness - #2520
perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness#2520cquil11 wants to merge 5 commits into
Conversation
db0bd06 to
dbaa77b
Compare
|
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 For PR verification, add the 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 如需进行 PR 验证，请为此 PR 添加 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 finished @cquil11's task in 1m 44s —— View job Review
LGTM - no blocking issues found Verified for the single changed file (
Note: I couldn't run a PyYAML parse directly (Python execution isn't in my allowed tools for this run), but the entry is structurally identical to the existing 5k+ lines of entries and introduces no unusual YAML constructs. |
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
There was a problem hiding this comment.
LGTM — approving despite the inline nit. The changelog entry itself is a single, well-formed addition consistent with existing schema and other entries in the file.
Extended reasoning...
Overview
This PR adds a single changelog entry to perf-changelog.yaml documenting a refresh of the dsv4-fp4-gb300-dynamo-sglang-agentic-agg AgentX harness. No other files are touched. The new entry follows the existing schema (config-keys, scenario-type, description, pr-link) and is consistent in structure with neighboring entries in the file.
Security risks
None — this is a metadata-only YAML changelog file with no executable code, credentials, or config that affects runtime behavior of the inference stack.
Level of scrutiny
Low. This is a mechanical, single-entry, single-file changelog addition with no logic changes. The one finding from this review round is a process/documentation nit (missing bilingual PR title/description per AGENTS.md), not a functional or correctness issue with the changelog content itself.
Other factors
A candidate concern that the changelog claims an 'AgentX harness refresh' with no corresponding code change was raised and ruled out by the verifier pass — changelog-only entries documenting external/out-of-repo harness refreshes appear elsewhere in this file's history, so this is not anomalous. Given the trivial scope and the nit being editorial rather than substantive, I'm comfortable approving.
| pr-link: https://github.com/SemiAnalysisAI/InferenceX/pull/2477 | ||
|
|
||
| - config-keys: | ||
| - dsv4-fp4-gb300-dynamo-sglang-agentic-agg | ||
| scenario-type: | ||
| - agentic-coding | ||
| description: | ||
| - "Refresh submission with up to date AgentX harness." |
There was a problem hiding this comment.
🟡 PR title and description are English-only, but AGENTS.md line 7 requires bilingual (English + Simplified Chinese) titles and descriptions for every PR. Please retitle to perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness / <中文标题> and add a ## 中文说明 section mirroring the Summary/Validation bullets.
Extended reasoning...
AGENTS.md (the repo's mandatory-reading doc, referenced from CLAUDE.md) states in its second paragraph, in bold, an explicit and unconditional requirement:
PR and GitHub-issue titles & descriptions must be bilingual — include a Simplified Chinese version in addition to English. Title format:
<English title> / <中文标题>. In the PR/issue body, follow the English content with its Chinese translation (e.g. a## 中文说明section mirroring the summary; ...)
PR #2520's title is perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness — there is no / <中文标题> suffix at all. The body only contains the English ## Summary and ## Validation sections; there is no accompanying ## 中文说明 (or any Chinese-language section) mirroring them.
This isn't a stale or ambiguous doc — the convention is actively followed elsewhere in this repo. For example, commit bb8e26d in this same repo's history uses the exact bilingual format the rule describes: [Klaud Cold] glm5.2-fp4-b300-sglang-agentic-mtp: ... / 新增 GLM-5.2 NVFP4 B300 SGLang EAGLE MTP AgentX 配置（黄金 AL 合成接受率）. So this is a live, enforced convention that PR #2520 simply didn't follow, not a dead rule.
Step-by-step proof:
- Open AGENTS.md at the repo root, line 7 (second bolded paragraph) — it mandates bilingual title format
<English title> / <中文标题>and a## 中文说明body section. - Open PR perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness #2520's title:
perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness. There is no/followed by Chinese characters anywhere in it. - Open PR perf(agentx): refresh dsv4-fp4-gb300-dynamo-sglang-agentic-agg harness #2520's description: it contains
## Summary(2 bullets) and## Validation(2 bullets), both entirely in English, with no## 中文说明or any Chinese text. - Compare against a compliant example already in this repo's git history (commit
bb8e26d), which does include the/ <中文标题>suffix — confirming the rule is actively enforced, not aspirational.
Impact: this is a documentation/process compliance gap only. The change itself (a one-entry addition to perf-changelog.yaml) is functionally fine and parses correctly, so this does not block the benchmark sweep or cause any runtime issue. The fix is purely editorial: append a Chinese title suffix and add a ## 中文说明 section translating the Summary/Validation bullets, following the terminology and style guidance also given in AGENTS.md (e.g. using the provided English/Chinese term mapping table).
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31150505447 |
…300-dynamo-sglang-agentic-agg
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31152169832 |
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31153422147 |
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=31153885543 |
1 participant
Summary
Validation