Poll: refactoring a chunk of legacy code

Interesting, this was an example that I got from Steven Johnson (the author of FFTW and a bunch of other high performance software). JuliaCon 2019 | Keynote: Professor Steven G. Johnson - YouTube, and I’ve seen this performance difference in practice when re-implimenting a variety of special functions. It’s possible that SciPy and SpecialFunctions.jl aren’t compiling it with the right flags (although that would be surprising), and I can’t think of any other possible cause of the difference.