Skip to content

feat(benchmarks): add agent-oriented math and transform benchmark cases#285

Open
haris-admin wants to merge 1 commit into
vercel-labs:mainfrom
haris-admin:agent-benchmarks
Open

feat(benchmarks): add agent-oriented math and transform benchmark cases#285
haris-admin wants to merge 1 commit into
vercel-labs:mainfrom
haris-admin:agent-benchmarks

Conversation

@haris-admin
Copy link
Copy Markdown

Summary

Adds three small agent-oriented benchmark cases under �enchmarks/agent/ exercising patterns that code agents commonly generate or repair.

Closes #8

Cases

  • gcd: Euclidean GCD with integer loop and branch behavior.
  • stats: Min, max, and sum over a fixed-size array.
  • token-count: Delimiter token scan over a byte span.

Changes

File Description
�enchmarks/agent/gcd.0 GCD benchmark source
�enchmarks/agent/stats.0 Stats benchmark source
�enchmarks/agent/token-count.0 Token count benchmark source
scripts/bench.mts Registered new cases in benchmark harness
docs/articles/benchmarks.md Documented agent-oriented cases

Verification

All three source files pass zero check on v0.1.4. Direct executable builds are blocked on Windows by the win32-x64 COFF backend limitation (BLD004), which existing benchmarks also hit. Recommended to verify full runtime correctness with pnpm run bench on a host with a complete direct backend.

Adds three small benchmark cases under benchmarks/agent/ exercising
patterns that code agents commonly generate or repair:

- gcd: Euclidean GCD with integer loop and branch behavior
- stats: Min, max, and sum over a fixed-size array
- token-count: Delimiter token scan over a byte span

Verified with zero check (v0.1.4) on all three files.

Closes vercel-labs#8
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.

Add small agent-oriented math and transform benchmark cases

1 participant