[AMD] retrigger dsv4-fp4-mi355x-atom benchmark sweep#1817
Conversation
…e image to atom0.1.4 - Enable --enable-tbo for ISL=1024/OSL=1024 at CONC>=1024 and ISL=8192/OSL=1024 at CONC>=256 - Update image to atom0.1.4_20260612 - Update ISL=8192 search-space to start at conc=4 and use DPA from conc=128 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…onc range - Pass --max-model-len to server using SERVE_MAX_MODEL_LEN - Add EVAL_ONLY path: compute eval context length via compute_eval_context_length - Extend conc-end to 8192 (isl=1024) and 4096 (isl=8192) in amd-master.yaml Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…sable max-model-len Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…m-seqs Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…e, enable max-model-len Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ap conc ranges Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ace description Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… remove stale entries Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…x mem frac Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…add TP4/TP8 conc lists for ISL8192 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…nch override Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…/TP8/DPA Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
|
Thanks for the contribution! For vLLM & SGLang, please ensure that your recipes is similar to the official vLLM recipes and/or the SGLang cookbook 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 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. |
|
Claude finished @Oseltamivir's task in 35s —— View job Review Complete
LGTM - no blocking issues found. The new entry is correctly appended at the end of |
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=27729964433 |
|
see unofficial run visualizer at https://inferencex.semianalysis.com/inference?unofficialRun=27730066598 |
|
/reuse-sweep-run 27676739575 |
2 participants
Summary
perf-changelog.yaml.dsv4-fp4-mi355x-atomwith the same image and search-space description.Validation
perf-changelog.yamlwith PyYAML.100644.Note
Low Risk
Append-only changelog metadata for CI orchestration; no runtime or config logic changes.
Overview
Retriggers the
dsv4-fp4-mi355x-atombenchmark sweep by appending a new block at the end ofperf-changelog.yaml. Sweep selection is driven by the changelog diff vsmain, so a fresh append is enough to kick off another run.The new entry is a duplicate of the existing PR #1717 changelog (same
config-keys, description bullets, andpr-link). It does not change.github/configs, launch scripts, or search-space YAML—only documents the re-run intent for reviewers.The described benchmark context (unchanged by this PR) is DeepSeek-V4 FP4 on MI355X ATOM: image
rocm/atom:…atom0.1.4_20260612, ISL=8192 search-space updates (TP8 conc 4–64, DPA conc 128–1024), and TBO at high concurrency.Reviewed by Cursor Bugbot for commit dbc4d69. Bugbot is set up for automated code reviews on this repo. Configure here.