[fanf2]

Apple’s M1 has a monstrously fast divide instruction.

There are some benchmarks here: https://ridiculousfish.com/blog/posts/benchmarking-libdivide...

I noticed it is very fast when I was experimenting with algorithms for unbiased bounded random numbers https://dotat.at/@/2022-04-20-really-divisionless.html - the fancy nearly and really divisionless algorithms had much less advantage on an Apple M1 than on an old Intel CPU.