What's fascinating is that it shows here that Hyperscan is not matching the same (!) as the other engines. A piece of information conveniently forgotten in the Intel piece.
Also note that the rust regex crate now uses the same SIMD algorithm(Teddy) invented as part of Hyperscan, see:
http://blog.burntsushi.net/ripgrep/ (also ripgrep is awesome, and you should use it.)
This is a surprisingly negative-sounding comment. We've been quite clear in our documentation and public language in the past about how 'all-matches' semantics differ from 'best match' semantics implemented by back-tracking engines, so we generally find a few more matches.
Given the comparison with RE2 here is under circumstances where RE2::Set is not returning any offset at all, I'm not sure what your point is about the "Intel piece".
Ripgrep is also awesome, and thanks (I guess) for mentioning Teddy, which nicely illustrates my favorite instruction (PSHUFB). There are still some bits of the Teddy algorithm that should be incorporated into ripgrep (the merging of literals when you have >8 literals is naive).
Also note that the rust regex crate now uses the same SIMD algorithm(Teddy) invented as part of Hyperscan, see: http://blog.burntsushi.net/ripgrep/ (also ripgrep is awesome, and you should use it.)