dsv4-fp4-mi355x-sglang: bump image to rocm720-mi35x-0363e6c-20260509-DSv4

[functionstackx]

Summary

• Bump dsv4-fp4-mi355x-sglang image to rocm/sgl-dev:rocm720-mi35x-0363e6c-20260509-DSv4 (one commit forward on amd/deepseek_v4)
• Add perf-changelog entry

Test plan

• Full sweep on dsv4-fp4-mi355x-sglang via the full-sweep-enabled label

🤖 Generated with Claude Code

[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://cookbook.sglang.io/
• https://docs.vllm.ai/projects/recipes/en/latest/

If it is not, please create a PR first before we can merge your 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/sgl-cookbook

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.

[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://cookbook.sglang.io/
• https://docs.vllm.ai/projects/recipes/en/latest/

If it is not, please create a PR first before we can merge your 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/sgl-cookbook

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.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25618904052
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25618904052

[claude]

LGTM, straightforward image tag bump with matching perf-changelog entry.

Extended reasoning...

Overview

This PR bumps the docker image for the dsv4-fp4-mi355x-sglang benchmark config from rocm720-mi35x-bfd32b6-20260507-DSv4 to rocm720-mi35x-0363e6c-20260509-DSv4 (one commit forward on the amd/deepseek_v4 branch) and adds a corresponding perf-changelog entry.

Security risks

None. This is a benchmark configuration file change that only affects which container image is pulled for sweeps on the mi355x runner. No code paths, auth, or permissions are touched.

Level of scrutiny

Low. This follows an established pattern (image tag bump) explicitly documented in the comment above the config block ("bumping sglang is just an image tag bump here"). The change is mechanical and self-contained.

Other factors

The PR is labeled full-sweep-enabled, so the actual validation will come from the full sweep run. The perf-changelog entry is properly formatted and matches the schema used by neighboring entries.

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25618907394
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25618907394

[github-actions]

see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=25635071835
see unofficial run visualizer at https://inferencex.semianalysis.com/evaluation?unofficialRun=25635071835