Hacktoberfest 2026: the issues maintainers tagged for October, open and beginner-friendly. Browse Hacktoberfest issues

Calibrate CSPA W=1 gate with paired hosted baselines

Open
  • Is something wrong?
#1,980 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
4/5
Estimated time
3-5 days
Newbie friendliness
48/100
Issue type
Feature
Clarity
Clearly specified
Activity status
Active
Tech stack
c
Domain
performance

Research direction

Start by locating the CSPA W=1 release benchmark and its performance-governor preflight, then inspect PR #1979 and the existing 2050 ms target. Run origin/main and the candidate PR on the same hosted runner with identical compiler, build options, fixtures, and invocation, alternating trial order; record the requested metadata and samples. Done means a reproducible paired report includes medians, CoV, paired delta, and whether the target transfers; invalid governor runs must be skipped, and any target change reviewed separately.

Written by the indexing model from the issue text.

Description

enhancement p2-medium

Discovery context

PR #1979's first stable hosted CSPA W=1 run measured a 3145.9 ms mean (3141.7 ms median, 12.24 ms standard deviation) against the existing 2050 ms target. Two later reruns skipped because the runner could not satisfy the performance-governor preflight. A local release-build comparison on a different schedutil host measured 4304.9 ms for origin/main and 4320.6 ms for the PR (0.4% difference), but that does not calibrate the hosted runner or explain the hosted result. There is not yet a same-host baseline pair, so the failure cannot be attributed to the PR or used to justify changing the target.

Scope

Run the CSPA W=1 release benchmark for origin/main and the candidate PR on the same hosted runner with identical compiler, build options, fixtures, and test invocation. Alternate trial order to reduce drift. Record CPU model/stepping, kernel, governor, compiler/version, build options, individual samples, medians, coefficient of variation, and paired delta. Preserve the existing performance-governor preflight; do not treat runs that fail it as measurements.

Acceptance criteria

  • A reproducible paired report shows the origin/main and PR CSPA W=1 results from the same host and identical build/test settings.
  • The report includes host/build metadata, per-trial samples, medians, CoV, and paired delta, and determines whether the 2050 ms target transfers to the current hosted runner.
  • Runs without the required performance governor are reported as skipped/invalid and never counted as evidence.
  • The target remains unchanged unless paired evidence supports a calibration; any target update is reviewed separately with the measurement provenance documented.
Dominant language
C
Stars
18
Forks
2
Avg merge
9h 48m
Merged PRs (30d)
165

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from semantic-reasoning/wirelog

All issues in semantic-reasoning/wirelog

Similar issues

More C issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.