Skip to content

ci+docs(benchmarks): run suites on Julia 1.12 + refresh page data#103

Merged
jack-champagne merged 2 commits into
mainfrom
benchmarks/julia-1.12-and-refresh
Jun 8, 2026
Merged

ci+docs(benchmarks): run suites on Julia 1.12 + refresh page data#103
jack-champagne merged 2 commits into
mainfrom
benchmarks/julia-1.12-and-refresh

Conversation

@jack-champagne

Copy link
Copy Markdown
Member

Standardizes the benchmark suites on Julia 1.12 (what most users run; convergence already used 1.12) and refreshes docs/src/benchmarks.md.

Changes

  • benchmark.yml + alloc-profile.yml: Julia 1.11 → 1.12.
  • docs/src/benchmarks.md: Environment → Julia 1.12; fixed dashboard cadence wording (per-release on v* tags, not "per-commit / every push to main" — corrected after the ci(benchmarks): publish dashboards on v* tags (fix never-publishing gate) #102 publish-gate fix).
  • Data tables refreshed with real 1.12 numbers (full-solve, evaluator micro, scaling) from this PR's own benchmark run — see follow-up commit. Replaces the prior dd0beb4 (Julia 1.11) snapshot.

🤖 Generated with Claude Code

…nce wording

- Bump benchmark.yml + alloc-profile.yml from Julia 1.11 to 1.12 so all three
  suites (timing, alloc, convergence) run on the same version users actually run
  (convergence was already 1.12). Updates the docs Environment table to 1.12.
- Fix now-incorrect dashboard cadence wording: after the #102 gate fix the series
  publish per-RELEASE on v* tags, not "per-commit" / "on every push to main".

Data tables (full-solve / micro / scaling) refreshed with real 1.12 numbers in a
follow-up commit on this branch, sourced from this PR's own benchmark run.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@codecov

codecov Bot commented Jun 8, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

…eba1ff)

Replaces the dd0beb4 (Julia 1.11) snapshot with values from this PR's own 1.12
benchmark run (artifact bench.json + JLD2, GH Actions ubuntu-latest, Julia
1.12.6, 4 threads):
- Full-solve, evaluator-micro, and scaling tables re-stamped.
- Scaling commentary rewritten: the fast-Ipopt cell shifted (now N=101/dim8 at
  0.66s vs MadNLP 15.29s), so the prose no longer names fixed "N=25" cells and
  notes the easy cell varies with the random draw.
- Environment: runner is 4 vCPU / 16 GB (was 2 vCPU / 7 GB); threads auto = 4.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@jack-champagne jack-champagne merged commit 204b9ee into main Jun 8, 2026
8 of 9 checks passed
@jack-champagne jack-champagne deleted the benchmarks/julia-1.12-and-refresh branch June 8, 2026 17:37
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant