fix(eval): measure CPU time in Top-K ERP resources - #631
Merged
Conversation
zzylol
marked this pull request as ready for review
September 11, 2026 02:23
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 Top-K evaluation adapter wrote elapsed wall time into ERP
*_cpu_secondsfields. Descheduling could therefore change a value presented as CPU cost, and existing catalogs could silently reuse those mislabeled measurements.Calibration now reads the Linux process CPU clock for update, merge, and readout operations. Catalog provenance declares that time basis, and selection/reproduction reject old or missing CPU provenance. Generation and end-to-end query latency remain wall time. The historical v2 report is explicitly marked; its numbers are not replaced by invented corrected results.
Before this PR: wall-clock samples populated CPU resource fields and were accepted by ERP selection.
After this PR: only newly calibrated process-CPU evidence is accepted. Unsupported platforms fail explicitly rather than substituting elapsed time.
Verification: three focused Rust tests passed, including real sketch update/merge/readout CPU measurement and rejection of missing timing provenance. This fixes the offline adapter in #602; it does not establish a production control-plane comparison or new performance benefits. Full release recalibration remains pending.