Skip to content

[Klaud Cold] Update DeepSeek-V4-Pro FP4 B200 Dynamo-SGLang AgentX agg image to v0.5.22-cu130 / [Klaud Cold] 将 DeepSeek-V4-Pro FP4 B200 Dynamo-SGLang AgentX 聚合式镜像更新至 v0.5.22-cu130 - #4203

Merged
adibarra merged 5 commits into
mainfrom
klaud/auto-934d5c1943090dd9-3af5706b64e41f45
Oct 21, 2026
Merged

adibarra merged 5 commits into
mainfrom
klaud/auto-934d5c1943090dd9-3af5706b64e41f45

Conversation

@Klaud-Cold

@Klaud-Cold Klaud-Cold commented Oct 16, 2026 •

Copy link
Copy Markdown
Collaborator

Goal: Update SGLang image from lmsysorg/sglang:nightly-dev-20260916-c9a8fba9 to lmsysorg/sglang:v0.5.22-cu130@sha256:e3a4f06001a29542fc4a8638536d9ffc5eb9c9942b08603c09b19a56eae4a1ae.
Baseline: 2026-09-22 · lmsysorg/sglang:nightly-dev-20260916-c9a8fba9
AgentX · PTP8x1/DTP8x0 · Mean latency · Sources: API 1, API 2

Concurrency Total tok/s/GPU ↑ Output tok/s/GPU ↑ TTFT ms ↓ TPOT ms ↓
1 2,459.2 18.11 860.89 3.3
4 3,908.82 27.09 648.44 3.87
8 6,750.42 49.81 650.74 4.54

Eval: N/A

中文

**目标:**将 SGLang 镜像从 lmsysorg/sglang:nightly-dev-20260916-c9a8fba9 更新为 lmsysorg/sglang:v0.5.22-cu130@sha256:e3a4f06001a29542fc4a8638536d9ffc5eb9c9942b08603c09b19a56eae4a1ae。
**基线:**2026-09-22 · lmsysorg/sglang:nightly-dev-20260916-c9a8fba9
AgentX · PTP8x1/DTP8x0 · 平均延迟 · 来源: API 1, API 2;数值及异常说明见上表。

Pin lmsysorg/sglang:v0.5.22-cu130 by digest and drop
SGLANG_OPT_SWA_SPLIT_LEAF_ON_INSERT, which no longer exists upstream.

将 B200 Dynamo-SGLang AgentX 聚合式配置的镜像按摘要固定为
lmsysorg/sglang:v0.5.22-cu130,并移除上游已删除的
SGLANG_OPT_SWA_SPLIT_LEAF_ON_INSERT 环境变量。

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@github-actions

Copy link
Copy Markdown
Contributor

Thanks for the contribution!

  • Review: If this PR changes files owned by someone other than a repository admin or @SemiAnalysisAI/core, ask one eligible CODEOWNER to complete the latest PR_REVIEW_CHECKLIST.md before contacting a core maintainer on Slack. Follow the template exactly, including As a PR reviewer and CODEOWNER, I have reviewed this and have, so sign-off verification triggers.
  • PR verification: Sweeps run only when the PR appends an inferencex-e2e/perf-changelog.yaml entry and carries exactly one primary label: full-sweep-fail-fast (strongly recommended; canary plus per-matrix fail-fast), full-sweep-enabled (canary; matrix jobs continue after a failure), or non-canary-full-sweep-enabled (no canary or fail-fast). The modifiers all-evals, evals-only, and agentx-fast require a primary label. On fork PRs, a maintainer applies the label. See sweep labels and reuse.
  • After merging: PR authors must ensure all GitHub Actions jobs pass. Transient failures often pass on rerun; see how to rerun failed jobs.
中文

感谢你的贡献!

  • **审阅:**如果 PR 修改的文件归属于仓库管理员及 @SemiAnalysisAI/core 之外的 CODEOWNER,请先联系一位有资格的 CODEOWNER 填写最新的 PR_REVIEW_CHECKLIST.md,再通过 Slack 联系核心维护者。必须严格遵循模板,并保留 As a PR reviewer and CODEOWNER, I have reviewed this and have,才能触发签核验证。
  • **PR 验证:**只有当 PR 在 inferencex-e2e/perf-changelog.yaml 末尾追加了条目,并且恰好带有一个主标签时,才会运行扫描:full-sweep-fail-fast(强烈推荐;canary 加逐矩阵 fail-fast)、full-sweep-enabled(有 canary;矩阵任务在失败后继续运行)或 non-canary-full-sweep-enabled(无 canary,也无 fail-fast)。修饰标签 all-evals、evals-only 和 agentx-fast 必须与主标签一起使用。fork PR 的标签由维护者添加。参见扫描标签与复用。
  • **合并后:**PR 作者必须确保所有 GitHub Actions 任务通过。临时性失败通常可以通过重新运行恢复;参见重新运行失败任务的说明。

@Klaud-Cold

Klaud-Cold commented Oct 16, 2026 •

Copy link
Copy Markdown
Collaborator Author

Initial attempt · Passed · Run 38948781345 / attempt 1 · 2026-10-16 20:09 UTC
lmsysorg/sglang:v0.5.22-cu130@sha256:e3a4f06001a29542fc4a8638536d9ffc5eb9c9942b08603c09b19a56eae4a1ae · ea0db03844a4 · AgentX · PTP8x1/DTP8x0 · Mean latency
Change: Measure branch head a3ae0a4: pin SGLang v0.5.22 (baa9befa, 1723 commits after c9a8fba9, CUDA 13.0.3, FlashInfer 0.7.1) by digest and drop SGLANG_OPT_SWA_SPLIT_LEAF_ON_INSERT, whose only reader left with the legacy SWA radix cache (sgl-project/sglang#40313); every recipe CLI flag and choice (DSPARK, flashinfer_mxfp4, FP4 indexer, HiCache write_through/direct/page_first_direct) exists at v0.5.22, SGLANG_NVFP4_CKPT_FP8_NEXTN_MOE stays default off, and no agg c8 eval was published for the 2026-09-22 baseline.

Concurrency Output tok/s/GPU ↑ TTFT ms ↓ TPOT ms ↓
1 18.21 (N/A) 960.58 (N/A) 3.22 (N/A)
4 27.14 (N/A) 752.65 (N/A) 3.75 (N/A)
8 50.49 (N/A) 700.05 (N/A) 4.53 (N/A)

Note: All rows: Δ N/A: dataset/scenario/statistic mismatch.

Eval Score ↑ Samples
gsm8k/em_strict · c8 97.35% (N/A) N/A/1,319 (old/new)

Note: All rows: Δ N/A: no matched eval baseline.

Next: Append the perf-changelog entry and run the full sweep on the updated image with full-sweep-fail-fast.

中文

初次尝试 · 已通过 · Run 38948781345 / attempt 1 · 2026-10-16 20:09 UTC
lmsysorg/sglang:v0.5.22-cu130@sha256:e3a4f06001a29542fc4a8638536d9ffc5eb9c9942b08603c09b19a56eae4a1ae · ea0db03844a4 · AgentX · PTP8x1/DTP8x0 · 平均延迟
**变更:**测量分支提交 a3ae0a4:按摘要固定 SGLang v0.5.22(baa9befa,比 c9a8fba9 新 1723 个提交,CUDA 13.0.3,FlashInfer 0.7.1),并移除 SGLANG_OPT_SWA_SPLIT_LEAF_ON_INSERT,其唯一读取方已随旧版 SWA radix cache 删除(sgl-project/sglang#40313);配方中所有 CLI 参数和取值(DSPARK、flashinfer_mxfp4、FP4 indexer、HiCache write_through/direct/page_first_direct)在 v0.5.22 中均存在,SGLANG_NVFP4_CKPT_FP8_NEXTN_MOE 保持默认关闭,2026-09-22 基线未发布聚合式 c8 评测。实测数值及异常说明见上表。
**下一步:**追加 perf-changelog 条目,并使用 full-sweep-fail-fast 在新镜像上运行完整扫描。

为 B200 聚合式 SGLang v0.5.22 镜像更新添加 perf-changelog 条目。

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@Klaud-Cold Klaud-Cold added the full-sweep-fail-fast Full sweep with canary gate; first failure cancels the rest of that matrix (recommended) label Oct 16, 2026
@github-actions

Copy link
Copy Markdown
Contributor

@Klaud-Cold

Copy link
Copy Markdown
Collaborator Author

Final full sweep · Passed · Run 38960572895 / attempt 1 · 2026-10-16 21:41 UTC
lmsysorg/sglang:v0.5.22-cu130@sha256:e3a4f06001a29542fc4a8638536d9ffc5eb9c9942b08603c09b19a56eae4a1ae · e82118bb200a · AgentX · PTP8x1/DTP8x0 · Mean latency
Change: Validate the complete updated-image family at this head.

Concurrency Output tok/s/GPU ↑ TTFT ms ↓ TPOT ms ↓
1 18.24 (N/A) 963 (N/A) 3.27 (N/A)
4 27.81 (N/A) 758.93 (N/A) 3.81 (N/A)
8 50.19 (N/A) 711.39 (N/A) 4.59 (N/A)

Note: All rows: Δ N/A: dataset/scenario/statistic mismatch.

Eval Score ↑ Samples
gsm8k/em_strict · c8 97.5% (N/A) N/A/1,319 (old/new)

Note: All rows: Δ N/A: no matched eval baseline.

Next: Mark ready for maintainer review.

中文

最终完整 sweep · 已通过 · Run 38960572895 / attempt 1 · 2026-10-16 21:41 UTC
lmsysorg/sglang:v0.5.22-cu130@sha256:e3a4f06001a29542fc4a8638536d9ffc5eb9c9942b08603c09b19a56eae4a1ae · e82118bb200a · AgentX · PTP8x1/DTP8x0 · 平均延迟
**变更:**验证此提交更新镜像后的完整配置族。实测数值及异常说明见上表。
**下一步:**标记为就绪,等待维护者审查。

@Klaud-Cold
Klaud-Cold marked this pull request as ready for review October 16, 2026 21:41
@Klaud-Cold
Klaud-Cold requested a review from a team October 16, 2026 21:41
@Klaud-Cold

Copy link
Copy Markdown
Collaborator Author

validated · Repairs: 0 · Runs: 38948781345, 38960500189, 38960572895
All owned runs ended. Full sweep verified; ready for review.

中文

validated · 修复次数:0 · 运行:38948781345, 38960500189, 38960572895
所有自有运行均已结束。完整 sweep 已验证;已就绪,等待审查。

@Klaud-Cold

Copy link
Copy Markdown
Collaborator Author

/use 38960572895

@github-actions

github-actions Bot commented Oct 16, 2026 •

Copy link
Copy Markdown
Contributor

@Klaud-Cold staged run 38960572895.

This run remains available across future /stage-results requests. Staging the same run ID again updates its staged data. Staging workflow

@claude claude Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code review found no issues

No high-confidence issues detected in this change.

Resolve perf-changelog.yaml by taking main's file and re-appending this PR's entry.

中文:合并 origin/main;perf-changelog.yaml 采用 main 版本,并在末尾重新追加本 PR 的条目。
Resolve conflicts: take main's refactored recipes and re-apply this PR's non-image edits; take main's perf-changelog.yaml and re-append this PR's entry.

中文:合并 origin/main;配方采用 main 重构后的版本并重新应用本 PR 的非镜像改动,perf-changelog.yaml 采用 main 版本并在末尾重新追加本 PR 的条目。
@adibarra

Copy link
Copy Markdown
Collaborator

/reuse-sweep-run 38960572895

@nvpohanh

Copy link
Copy Markdown
Collaborator

[by Claude Code] The SGLang cookbook update of this PR's AgentX config (dsv4-fp4-b200-dynamo-sglang-agentic-agg) will happen in sgl-project/sglang#46807, in the Agentic Long-Context Deployment section of the DeepSeek-V4 cookbook page. The config is already live in that section; sgl-project/sglang#46807 refreshes it with the pin below.

The commands mirror the launch measured in run 38960572895, pinned by this PR's /reuse-sweep-run comment.

@nvpohanh

Copy link
Copy Markdown
Collaborator

[by Claude Code] sgl-project/sglang#46807 has been merged. This PR's AgentX config (dsv4-fp4-b200-dynamo-sglang-agentic-agg) is now in the Agentic Long-Context Deployment section of the SGLang DeepSeek-V4 cookbook: https://docs.sglang.io/cookbook/autoregressive/DeepSeek/DeepSeek-V4#agentic-long-context-deployment

The commands mirror the launch measured in run 38960572895, pinned by this PR's /reuse-sweep-run comment.

@nvpohanh

Copy link
Copy Markdown
Collaborator

[by Claude Code] The SGLang cookbook update of this PR's AgentX config (dsv4-fp4-b200-dynamo-sglang-agentic-agg) will happen in sgl-project/sglang#46948, in the Agentic Long-Context Deployment section of the DeepSeek-V4 cookbook page. The config has been live in that section since sgl-project/sglang#46807; sgl-project/sglang#46948 refreshes it with the pin below.

The commands mirror the launch measured in run 38960572895, pinned by this PR's /reuse-sweep-run comment.

@nvpohanh

Copy link
Copy Markdown
Collaborator

[by Claude Code] sgl-project/sglang#46948 has been merged. This PR's AgentX config (dsv4-fp4-b200-dynamo-sglang-agentic-agg) is now in the Agentic Long-Context Deployment section of the SGLang DeepSeek-V4 cookbook: https://docs.sglang.io/cookbook/autoregressive/DeepSeek/DeepSeek-V4#agentic-long-context-deployment

The commands mirror the launch measured in run 38960572895, pinned by this PR's /reuse-sweep-run comment.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

full-sweep-fail-fast Full sweep with canary gate; first failure cancels the rest of that matrix (recommended)

Projects

Status: Done

Development

Successfully merging this pull request may close these issues.

4 participants