Compare retained installed timing observations

[kantorcodes]

Compare two retained installed builds using the existing timing checks in a fixed previous/current/current/previous order. Verify the selected artifacts and unchanged benchmark inputs, then record bounded observations without adding requests or changing warmup and limits.

Validation: seven diagnostic-contract tests and repository style checks pass. Both retained artifacts match their pinned manifests and executable hashes. The completed comparison produced four valid observations: three passed all timing checks; the final observation of the previous build exceeded the existing concurrency limit. Every observation completed all requests without errors.

The result demonstrates variation within an unchanged retained build. It does not establish a source-specific cause or runtime acceptance, and the earlier failed gate remains recorded.

[coderabbitai]

Important

Draft PR not reviewed

Draft PRs are not automatically reviewed by default.

• Trigger a manual review

To automatically review draft PRs, update your CodeRabbit configuration:

reviews:
  auto_review:
    drafts: true

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gitar-bot]

CI failed: CI failure in the installed-timing-comparison workflow due to a non-zero exit code from the comparison script.

Overview

1 log analyzed showing a failure in the installed-timing-observations comparison step where the script exited with code 1.

Failures

Installed Timing Comparison Script Failure (confidence: medium)

• Type: test
• Affected jobs: 105461712133
• Related to change: yes
• Root cause: The script 'scripts/ci/compare_installed_timing.py' returned a non-zero exit code (1), indicating a failure or assertion error when comparing the retained installed timing observations.
• Suggested fix: Examine the detailed execution logs of the timing comparison script or run it locally to troubleshoot the failure.

Summary

• Change-related failures: 1 timing comparison script failure
• Infrastructure/flaky failures: None
• Recommended action: Review the output of the 'compare_installed_timing.py' script execution to fix the assertion or error causing exit code 1.

Code Review ✅ No issues found

🟡 Medium risk · Adds a macOS CI workflow that installs artifacts and runs timing comparisons

Compares two retained installed builds using existing timing checks in a fixed previous/current/current/previous order, verifying artifact selection and unchanged benchmark inputs while recording bounded worker and timing observations. Seven diagnostic-contract tests pass, style checks pass, and both retained wheels match their pinned manifests and executable hashes. No issues found.

Review coverage

📋 Rules No rules evaluated

🧪 Functional validation Not enabled · Set up

Options ✅ Auto-apply

✅ Auto-apply is on → Gitar will commit updates to this branch.
Display: compact → Counting what did not apply, without listing it.

Comment with these commands to change the behavior for this request:

✅ Auto-apply	Compact
gitar auto-apply:off	gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Powered by Gitar — free for open source}