On Mon, Oct 17, 2022 at 11:27:17AM -0700, Eric Biggers wrote: > On Sun, Oct 16, 2022 at 08:37:53PM -0600, Jason A. Donenfeld wrote: > > In order to be efficient, we implement a kernel-specific variant of > > Daniel Lemire's algorithm from "Fast Random Integer Generation in an > > Interval", linked below. The kernel's variant takes advantage of > > constant folding to avoid divisions entirely in the vast majority of > > cases, works on both 32-bit and 64-bit architectures, and requests a > > minimal amount of bytes from the RNG. > > > > Link: https://arxiv.org/pdf/1805.10941.pdf > > Thanks for doing this! Your code looks correct, but it was hard for me to > understand until I read the paper that is linked to. Could you include a brief > comment in the code that explains the algorithm? Also, though the code looks > correct, I assume that you've also explicitly tested that each of the four code > paths produce uniform random numbers as intended? Yes, I've tested those, and they work. (Threw a lot of cores and ram at it.) I could include a comment, sure. What do you have in mind? A line-by-line thing, or just a short blurb at the top of the function? Jason
