This is additional details on how I checked the performance of the Oxlint beta release, so that you can reproduce and check the performance for yourself.
The computer I used for benchmarking is an Apple M1 MacBook Air with 16GB of RAM. I closed as many programs as possible prior to benchmarking, leaving mainly just Ghostty running, which I used as a terminal.
The oxlint versions I used were:
- GA / alpha: The
oxlint-darwin-x64executable from the v0.0.20 release: https://github.com/oxc-project/oxc/releases/tag/oxlint_v0.0.20, this executable was namedoxlint-alpha. - Beta: Built oxlint using
cargo build --releaseat f707d1f5fe7fa88c516f6373b87920eb5873f8f4, and named the executableoxlint-beta.
For benchmarking, I used https://github.com/sharkdp/hyperfine at version 1.19.0. The commands used for benchmarking each version were:
- GA:
hyperfine -i -N --warmup 5 --runs 100 './oxlint-alpha <DIRECTORY>' - Beta:
hyperfine -i -N --warmup 5 --runs 100 './oxlint-beta <DIRECTORY>'
There were 5 warmup runs for each repository, and then 100 runs. The lint time recorded was the average across all runs.
The repositories used for benchmarking were:
- https://github.com/microsoft/vscode at 3e749796fa7e9d82be625e7f077b81a6741b06e7
- https://github.com/elastic/kibana at 3cb6a63807766758f5c1a7d827ead164936c0439
- https://github.com/vuejs/core at 29de6f8b0bb1a604f247b0712daac29e93aa6f3e
- https://github.com/vitest-dev/vitest at 3fdee3da8b281da7404ab78f11b4984ec493cd54