Skip to content

[Doc] cookbook: Refresh SemiAnalysis AgentX launches to InferenceX main ce35bc93 - #48857

Merged
nvpohanh merged 1 commit into
sgl-project:mainfrom
nvpohanh:pohanh/cookbook-agentx-refresh-1105c
Nov 6, 2026
Merged

nvpohanh merged 1 commit into
sgl-project:mainfrom
nvpohanh:pohanh/cookbook-agentx-refresh-1105c

Conversation

@nvpohanh

@nvpohanh nvpohanh commented Nov 6, 2026 •

Copy link
Copy Markdown
Collaborator

[by Claude Code]

Motivation

The cookbook's Agentic Long-Context Deployment sections (#42726, #43083, #43799, #43983, #44065, #44228, #44654, #44712, #45039, #45111, #45311, #45513, #45749, #46005, #46506, #46547, #46807, #46948, #47027, #47214, #47464, #47929, #48392, #48551, #48634) mirror SemiAnalysis InferenceX AgentX submissions as raw launch commands, each linked to the run that measured it. The configurations are mirrored from SemiAnalysis submissions. This PR:

  1. Syncs every page to InferenceX main (ce35bc93). New or updated merged lanes come from their dashboard runs:

    The Kimi-K3.1-Pro page gets GB200 and B200 hardware options (GB300, GB200, B300, B200, plus MI355X). The MI355X data stays the hand-maintained file from [Doc] cookbook: Add Kimi-K3.1-Pro deployment recipes #48413.

  2. Overlays four open InferenceX PRs:

Source

InferenceX main (ce35bc93), with these InferenceX PRs overlaid:

PR Page Lanes Pinned to
#4853 Qwen3.5 GB300 Dynamo+SGLang on the official SGLang nightly with NVIDIA megakernels Qwen3.5 qwen3.5-fp4-gb300-dynamo-sglang-agentic-agg, -disagg no /use yet: in progress at head 2c6b0317
#4850 Qwen3.8-Flash-Next B200 SGLang AgentX update Qwen3.8-Flash-Next qwen3.8next-fp4-b200-sglang-agentic-mtp no /use yet: in progress at head fcf2a217
#4787 Qwen3.8-Flash-Next GB300 SGLang AgentX update Qwen3.8-Flash-Next qwen3.8next-fp4-gb300-sglang-agentic-mtp no /use yet: in progress at head 1aa39c5b
#4757 Muse Spark Open GB300 Dynamo+SGLang: retuned agg points, new TP8 disagg points Muse Spark Open musespark-fp4-gb300-dynamo-sglang-agentic-dflash-agg, -disagg no /use yet: in progress at head 7caf5718

"In progress" lanes are rendered at the PR head and labeled "In progress in SA InferenceX AgentX". A later sync pins them once the PR gets a /use run or merges.

What this PR does

Page Lanes Deployments Points Hardware
Kimi-K3.1-Pro 6 (was 3) 17 (was 10) 41 (was 20) GB300, GB200, B300, B200 (+ MI355X, hand-maintained)
Qwen3.5 16 (was 18) 58 (was 52) 231 (was 222) GB300, GB200, B300, B200, H200, H100
Qwen3.8-Flash-Next 4 10 (was 8) 26 (was 25) GB300, B300, B200, H200
MiniMax-M3.1 6 (was 8) 36 (was 40) 108 (was 119) GB300, GB200, B300, B200
Muse Spark Open 7 15 61 GB300, GB200, B300, B200, H200
DeepSeek-V4.1 10 22 67 GB300, GB200, B300, B200, H200, H100
GLM-5.2 14 34 75 GB300, GB200, B300, B200, H200
GLM-5.3 1 1 10 B300
DeepSeek-V4 unchanged
  • One new mirrored setting, on the Qwen3.5 GB300 disagg decode workers (InferenceX #4853), kept as submitted: SGLANG_PLB_SCATTER_DEFER=1. The decode worker queues incoming PD KV scatters and runs them between forward passes instead of on arrival (SGLang default off).
  • Component fix (_agentx.jsx): a point can now carry its own config files (point.files) when they differ from the first point of its deployment. Before, every point showed the first point's files. In the Qwen3.5 GB300 3P1D PP4/DEP8 deployment, c720 runs 4 Dynamo frontends behind nginx and no --router-policy-config. It would otherwise have shown c704's 2-frontend nginx.conf and a stray router_policy_config.yaml tab. No other point on any page is affected.
  • The Muse Spark Open, DeepSeek-V4.1 and GLM-5.2 data change only in provenance (the re-pinned or newly merged lanes above; commands unchanged), and GLM-5.3 only in the shared source block (the InferenceX main SHA and the overlay list).

Validation

  • Browser checks: every page this PR changes was rendered in a headless browser (mint dev, Mintlify as in CI).

    • Every hardware / checkpoint / deployment / concurrency / KV tier / router combination renders exactly the decoded commands: 3,408 command sets, 0 mismatches, 0 unseen, no browser errors. (On Kimi-K3.1-Pro the 13 MI355X combinations come from the hand-maintained file, which the sync does not generate, so they have no decoded reference and were not compared.)
    • Every tab was clicked, and all 197 "Copy link" URLs were reopened.
    • The section links from the Deployment panels and the copy buttons were also checked.
  • Decode and KV-flag checks:

    • The component's own decoder reproduces every generated plan (3,408 router × KV-tier variants).
    • Every KV flag exists in its image's SGLang commit (229 image × flag pairs).
  • Arg-parse: all 296 commands not validated in an earlier round parse. Each image was parsed with itself; both platforms of each image are built from the same commit:

    Image Commands Build commit
    nightly-dev-cu13-20261106-858073e4 159 858073e4
    dev-kimi-k31-pro 110 586355a8
    v0.5.23-cu130 27 297e54fd
  • node scripts/check_cookbook_configs.mjs, mint broken-links --check-anchors --check-redirects, mint validate and pre-commit run --all-files (in lmsysorg/sglang:dev-cu13) pass.

🤖 Generated with Claude Code


CI States

Latest PR Test (Base): ✅ Run #42188658741
Latest PR Test (Extra): ❌ Run #42188658641
Latest PR Test (AMD ROCm 10): ➖ No AMD PR run found for this commit.

Tip

Base CI only covers the core cases, and its workload may not exercise your change. Use /rerun-test to run the tests your change touches, or copy this prompt for an agent:

Find the registered tests this PR affects and run them with /rerun-test, following https://github.com/sgl-project/sglang/blob/main/.agents/skills/rerun-test/SKILL.md

…in ce35bc93

Mirror the SemiAnalysis InferenceX AgentX SGLang / Dynamo+SGLang lanes at
InferenceX main ce35bc93 (adds the Kimi-K3.1-Pro B200 and GB200
Dynamo+SGLang lanes from InferenceX sgl-project#4801 / sgl-project#4804; sgl-project#4811, sgl-project#4707, sgl-project#4670 and
sgl-project#4673 merged), with InferenceX PRs sgl-project#4853 (Qwen3.5 GB300 Dynamo+SGLang on
the SGLang nightly with NVIDIA megakernels), sgl-project#4850 / sgl-project#4787
(Qwen3.8-Flash-Next B200 / GB300 SGLang) and sgl-project#4757 (Muse Spark Open GB300
Dynamo+SGLang) overlaid.

The AgentX component now takes per-point config files (point.files), for
points whose nginx.conf / router policy differ from the first point of
their deployment (Qwen3.5 GB300 3P1D c720).

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

@claude claude Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

Claude Code Review

This pull request is from a fork — automated review is disabled. A repository maintainer can comment @claude review to run a one-time review.

@mintlify

mintlify Bot commented Nov 6, 2026 •

Copy link
Copy Markdown
Contributor

✅ Preview approved

@nvpohanh approved a Mintlify preview for commit c6c2e87.

@mintlify

mintlify Bot commented Nov 6, 2026 •

Copy link
Copy Markdown
Contributor

Preview deployment for your docs. Learn more about Mintlify Previews.

Project Status Preview Updated
sglang-doc 🟢 Ready View Preview Nov 6, 2026, 8:44 AM

💡 Tip: Enable Automations to automatically generate PRs for you.

@nvpohanh

nvpohanh commented Nov 6, 2026

Copy link
Copy Markdown
Collaborator Author

preview looks good. merging

This branch was successfully deployed

1 active deployment
staging - docs — c6c2e87e Deployed Nov 6, 2026 by mintlify[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

deepseek documentation Improvements or additions to documentation

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant