I love how easy rayon makes it to introduce a thread-pool to parallelize a problem over in #rust .

3-4 hours of work resulted in -84% running time on one of our benchmarks.

Now it feels stupid that I haven't done this before.

@alexanderkjall to be fair, I had a university course where we used OpenMP in C and that was pretty nice too (`#pragma omp parallel for` go brrr) ... but yes, that doesn't come with any of the memory / race / thread safety guarantees that Rust + rayon comes with