perf(linter): use parallel iterator directly instead of iter and parallel bridge

[camchenry]

Experimenting with some threading optimization, I noticed that ParallelBridge says that using this can be less efficient than just using par_iter directly:

https://docs.rs/rayon/latest/rayon/iter/trait.ParallelBridge.html

This creates a “bridge” from a sequential iterator to a parallel one, by distributing its items across the Rayon thread pool. This has the advantage of being able to parallelize just about anything, but the resulting ParallelIterator can be less efficient than if you started with par_iter instead. However, it can still be useful for iterators that are difficult to parallelize by other means, like channels or file or network I/O.

May not reflect in codspeed since this needs multiple files to be noticeable. Going to benchmark locally first.

[camchenry]

• perf(linter): use parallel iterator directly instead of iter and parallel bridge #8831 👈 (View in Graphite)
• main

---

How to use the Graphite Merge Queue

Add either label to this PR to merge it via the merge queue:

• 0-merge - adds this PR to the back of the merge queue
• hotfix - for urgent hot fixes, skip the queue and merge this PR next

You must have a Graphite account in order to use the merge queue. Sign up using this link.

An organization admin has enabled the Graphite Merge Queue in this repository.

Please do not merge from GitHub as this will restart CI on PRs being processed by the merge queue.

This stack of pull requests is managed by Graphite. Learn more about stacking.

[codspeed-hq]

CodSpeed Performance Report

Merging #8831 will not alter performance

_{Comparing 02-01-perf_linter_use_parallel_iterator_directly_instead_of_iter_and_parallel_bridge (c5ca0ba) with main (0568210)}

Summary

✅ 33 untouched benchmarks

[camchenry]

This didn't end up making any appreciable difference when benchmarking, so closing

[Boshen]

Let me take a closer look. Codspeed does not bench on multiple files.

[camchenry]

@Boshen I did benchmark it locally on vscode but didn't see any difference, sometimes it was slower, sometimes it was faster. Statistically it was all just noise. Theoretically it seems like this should be better based on the docs so maybe it's still valid to merge