A Quick Look at Xoshiro256**

submited by
Style Pass
2023-03-16 18:30:06

On May 4, David Blackman and Sebastiano Vigna announced new members of the Xoroshiro family and a new test for random number generators (based on the z9 test from gjrand) that their previous work fails, all described in a new paper. They claim to have now developed an “all-purpose, rock-solid generator”. In this post, having had less than a day to review their work, I'll present a few preliminary thoughts on this news, mostly looking at their best new generator, xoshiro256**.

One good thing about Blackman & Vigna's new paper is that they finally seem to be conceding that Xoroshiro128+ is a flawed generator. They have also updated the constants used in Xoroshiro128+ to create a new version of this PRNG (somewhat confusingly labeled “Version 1.0”), with different output, that gives some improvements to its performance in their new statistical test (which it still fails). Vigna appears to have also edited the Wikipedia page for Xoroshiro128+ to reflect the updated comments and caveats in the updated source and delete some of the other content from the page.

One thing Vigna doesn't seem to have conceded is that his interleaving-only approach to testing with TestU01 is insufficient, and that when a more comprehensive testing approach is used, Xoroshiro128+ fails TestU01.

Leave a Comment