Build the freeze job's benchmark optimized - #505
Merged
Conversation
The KMeans parity job's artifact is what the freeze step promotes to headline_result_v2.json, so it decides the published numbers. It was still building Flow at -O0 against scikit-learn's optimized wheels, which is the comparison the -O3 move corrected in flow.yml but not here. The effect was visible the moment #504 merged. main froze 16 of 19 with LinearSVC on digits at 510 ms and DecisionTree at 27.8 ms, against 155 ms and 9.9 ms for the same commit in the optimized benchmark job, and against 19 of 19 on four other machines. flow.yml:111 stays at -O0 on purpose. That job runs the test suite rather than the benchmark, and compiles faster for it. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
godofecht
added a commit
that referenced
this pull request
Sep 7, 2026
The KMeans parity job's artifact is what the freeze step promotes to headline_result_v2.json, so it decides the published numbers. It was still building Flow at -O0 against scikit-learn's optimized wheels, which is the comparison the -O3 move corrected in flow.yml but not here. The effect was visible the moment #504 merged. main froze 16 of 19 with LinearSVC on digits at 510 ms and DecisionTree at 27.8 ms, against 155 ms and 9.9 ms for the same commit in the optimized benchmark job, and against 19 of 19 on four other machines. flow.yml:111 stays at -O0 on purpose. That job runs the test suite rather than the benchmark, and compiles faster for it.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The
Canonical 19/19 KMeans parityjob produces the artifact that the freeze step promotes toheadline_result_v2.json, so it decides the published numbers. It was still building Flow at-O0against scikit-learn's optimized wheels.That is the same mismatch #504 corrected in
flow.yml, missed here because this workflow carries its own env block.The effect showed up the moment #504 merged.
mainfroze 16 of 19:Four other measurements of that commit read 19 of 19.
flow.yml:111stays at-O0deliberately: that job runs the test suite rather than the benchmark, and compiles faster for it.