We are considering using MD5 for user id + salt hashing for randomizing users for A/B testing. Should blake3 or xxhash work as a replacement for that use case?
This article is focussed on performance of large data sets.
Unless you have gigabytes of user IDs to hash on a regular basis, it doesn't matter what hash you use aside from achieving the randomness you desire, which I imagine almost any function, even x%2==0, would achieve.
Unless you have gigabytes of user IDs to hash on a regular basis, it doesn't matter what hash you use aside from achieving the randomness you desire, which I imagine almost any function, even x%2==0, would achieve.