Surprisingly, seems like it's reasonably good! There's a section in this blog post about characterizing it and comparing it to a hardware implementation: https:

Surprisingly, seems like it's reasonably good! There's a section in this blog post about characterizing it and comparing it to a hardware implementation: https://www.bunniestudios.com/blog/?p=6521 https://www.bunniestudios.com/blog/?p=6521 Since it's the same algorithm, most of the tricks appear to make it through the language conversion, so it's reasonably good even on a completely different architecture.