[lifthrasiir]

Just in case, here "Unicode routines" refer to Unicode encoding routines, not other more complex things like normalization. They are pretty regular but their automata would be relatively simple. (By the way, SIMD-accelerated regular expression parsing is indeed a thing [1], if anyone wondered.)

[1] https://www.microsoft.com/en-us/research/wp-content/uploads/...

[wooosh]

Intel has an implementation of this technique here as well:

https://github.com/intel/hyperscan

[nextaccountic]

ripgrep also uses simd https://news.ycombinator.com/item?id=19271497