Update DSv4 GB300 Dynamo SGLang MTP image

[Fridge003]

Summary

• Update dsv4-fp4-gb300-dynamo-sglang-mtp to lmsysorg/sglang:nightly-dev-20260527-14f81a67, following [GB300][SGLang] Enable W4A4 megamoe and bump SGLang image for dsv4-fp4-gb300-dynamo-sglang-mtp #1378

---

Note

Low Risk
Config-only image tag and changelog update with no application or auth logic changes.

Overview
Bumps the container image for dsv4-fp4-gb300-dynamo-sglang-mtp from lmsysorg/sglang:nightly-dev-cu13-20260509-9ee83034 to lmsysorg/sglang:nightly-dev-20260527-14f81a67 in .github/configs/nvidia-master.yaml.

Adds a matching perf-changelog.yaml entry for that config key documenting the image change (PR #1637).

^{Reviewed by Cursor Bugbot for commit 6ce9969. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook

• https://docs.sglang.io/cookbook/intro
• https://recipes.vllm.ai/

If it is not, please create a PR first before we can merge your single node PR into the master branch. Let's ensure that the documentation is first class such that the entire ML community can benefit from your hard work! Thank you

• https://github.com/vllm-project/recipes
• https://github.com/sgl-project/sglang/tree/main/docs_new

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. If re-running failed jobs is attempted, PR authors are responsible for ensuring it passes. See GitHub's docs on re-running failed jobs: https://docs.github.com/en/actions/how-tos/manage-workflow-runs/re-run-workflows-and-jobs#re-running-failed-jobs-in-a-workflow

As a rule of thumb, generally, PR authors should request a review & get a PR approval from the respective companies' CODEOWNERS before requesting a review from core maintainers.

If additional help is needed, PR authors can reach out to core maintainers over Slack.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit a2c5d68. Configure here.}

[cursor]

MTP image missing cu13 tag for GB300 hardware

High Severity

The image was changed from nightly-dev-cu13-20260509-9ee83034 (CUDA 13) to nightly-dev-20260527-14f81a67 (CUDA 12). SGLang publishes separate nightly-dev-* and nightly-dev-cu13-* variants for CUDA 12 and CUDA 13 respectively. The sibling config dsv4-fp4-gb300-dynamo-sglang on the same gb300-cw runner correctly uses nightly-dev-cu13-20260520-425dffbd. Running a CUDA 12 image on GB300 hardware requiring CUDA 13 will likely cause runtime failures.

 

^{Reviewed by Cursor Bugbot for commit a2c5d68. Configure here.}

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26811768166
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26811768166

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26842771985
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26842771985

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26842780129
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26842780129

[Oseltamivir]

lgtm

[Oseltamivir]

/reuse-sweep-run

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=26856602694
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=26856602694