Basic concurrent checking

[MichaReiser]

Summary

This PR makes workspace.check to check the files concurrently.

This PR doesn't explore more advanced scheduling. For example, we could try to not only schedule the workspace files
but also the dependencies of those files. This would give us better performance where a project has few first-party modules but depends on many third-party modules.

The main downside of concurrent checking is that the logs become harder to read, especially when using -vvv because the output of different threads is intertwined. I don't have a good solution for this yet other than using RAYON_NUM_THREADS=1 to explicitly fall-back to a single thread.

Test Plan

Checking black goes down from 108ms to 34ms (12 cores)

Main

black (cold)
Benchmark 1: knot
  Time (mean ± σ):     125.2 ms ±   3.8 ms    [User: 89.3 ms, System: 35.7 ms]
  Range (min … max):   117.7 ms … 133.3 ms    24 runs
 
  Warning: Ignoring non-zero exit code.
 
jinja (cold)
Benchmark 1: knot
  Time (mean ± σ):     139.1 ms ±   9.2 ms    [User: 100.1 ms, System: 38.8 ms]
  Range (min … max):   125.0 ms … 155.7 ms    22 runs
 
  Warning: Ignoring non-zero exit code.
 
isort (cold)
Benchmark 1: knot
  Time (mean ± σ):      85.2 ms ±   2.3 ms    [User: 60.1 ms, System: 25.0 ms]
  Range (min … max):    80.6 ms …  89.1 ms    32 runs
 
  Warning: Ignoring non-zero exit code.

Concurrent

uv run benchmark  --min-runs=3 --knot
black (cold)
Benchmark 1: knot
  Time (mean ± σ):      36.1 ms ±   1.1 ms    [User: 125.6 ms, System: 71.7 ms]
  Range (min … max):    33.3 ms …  39.4 ms    77 runs
 
  Warning: Ignoring non-zero exit code.
 
jinja (cold)
Benchmark 1: knot
  Time (mean ± σ):      32.6 ms ±   1.5 ms    [User: 140.8 ms, System: 88.7 ms]
  Range (min … max):    29.5 ms …  36.4 ms    88 runs
 
  Warning: Ignoring non-zero exit code.
 
isort (cold)
Benchmark 1: knot
  Time (mean ± σ):      30.7 ms ±   1.2 ms    [User: 82.4 ms, System: 46.7 ms]
  Range (min … max):    28.4 ms …  34.4 ms    89 runs
 
  Warning: Ignoring non-zero exit code.

[codspeed-hq]

CodSpeed Performance Report

Merging #13049 will not alter performance

_{Comparing concurrent=checking (acbed7c) with main (1f2cb09)}

Summary

✅ 32 untouched benchmarks

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.

[carljm]

Awesome!