Skip to content

refactor(benchmarks): group artifacts by benchmark family - #567

Merged
aviggiano merged 5 commits into
mainfrom
refactor/513-benchmark-layout
Aug 13, 2026
Merged

aviggiano merged 5 commits into
mainfrom
refactor/513-benchmark-layout

Conversation

@aviggiano

@aviggiano aviggiano commented Aug 13, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

  • organize benchmarks under evmbench, ultrafuzzbench, and scfuzzbench roots
  • move cohorts, policies, labels, history, and archived results to their owning benchmark family
  • update runtime, CLI, CI publication, cleanup, documentation, and test path consumers
  • require canonical family-scoped policy paths with no root-layout fallbacks

Verification

  • pnpm -w format:check
  • pnpm -w lint
  • pnpm -w typecheck
  • pnpm -w benchmark:check:prebuilt
  • pnpm -w docs:check
  • pnpm -w test:ci-scripts (66 tests)
  • pnpm --filter @ultrafuzz/evals test (392 tests)
  • focused publication, launch, history, and manifest tests (64 tests)

Closes #513

Greptile Summary

The PR reorganizes benchmark manifests, labels, history, and archived results under benchmark-family directories and updates their runtime, CLI, CI, documentation, and test consumers.

  • Moves EVMBench, UltrafuzzBench, and SCFuzzBench artifacts beneath family-specific roots.
  • Updates benchmark preparation, Modal execution, history publication, workflow filters, and CLI defaults to use the new paths.
  • Migrates checked-in history publication URLs and public result archives.

Confidence Score: 4/5

The PR is not yet safe to merge because cleanup and history publication for already-running pre-refactor candidates still fail on missing legacy policy paths.

The existing thread reply claims the cleanup path was fixed with a shared current-or-legacy resolver, but current HEAD hardcodes only family-scoped manifests in launch validation, publication preparation, and history scope validation, providing a concrete counterexample for legacy candidates.

Files Needing Attention: scripts/ci/validate-modal-benchmark-launch.mjs, scripts/ci/prepare-eval-history-publication.mjs, packages/evals/src/history.ts

Important Files Changed

Filename Overview
scripts/ci/validate-modal-benchmark-launch.mjs Switches launch and cleanup policy loading to canonical family-scoped manifests, but the previously reported legacy-candidate cleanup failure remains at HEAD.
scripts/ci/prepare-eval-history-publication.mjs Uses the new family-scoped policy files consistently, while still rejecting legacy candidate checkouts during publication preparation.
packages/evals/src/history.ts Updates public-generation validation to family-scoped manifests but provides no legacy candidate resolution for in-flight pre-refactor runs.
packages/modal/src/public-worker.ts Updates public worker cohort, lane, and ground-truth paths consistently with the new benchmark layout.
scripts/ci/publish-eval-history-cas.mjs Moves the CAS publisher’s permitted history path to the new UltrafuzzBench location.

Fix All in Greploop

Reviews (4): Last reviewed commit: "docs(evals): use canonical SCFuzzBench p..." | Re-trigger Greptile

Context used:

@aviggiano
aviggiano marked this pull request as ready for review August 13, 2026 14:25
@aviggiano

Copy link
Copy Markdown
Collaborator Author

@aviggiano please review

Comment thread scripts/ci/validate-modal-benchmark-launch.mjs
@aviggiano
aviggiano merged commit 056d1e4 into main Aug 13, 2026
9 checks passed
@aviggiano
aviggiano deleted the refactor/513-benchmark-layout branch August 13, 2026 19:46
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.

Refactor benchmarks/ directory

1 participant