Re:New Level 7 Hash Function: Prvhash
by aleksv on Monday August 17, 2020 @09:48PM
(#24194483)
PRVHASH is a hash function that generates a pseudo-random number sequence derived from the message. Resulting hashes closely follow normal distribution of bit frequency. PRVHASH is conceptually similar to keccak scheme, but is a completely different implementation of this concept.
PRVHASH can generate 32- to unlimited-bit hashes, yielding hashes of roughly equal quality independent of the chosen hash length. PRVHASH is based on 64-bit math. Hashes beyond 256-bits still require extensive testing, but, for example, any 32-bit element extracted from 512- or 2048-bit resulting hash is as collision resistant as just a 32-bit hash. The use of the function beyond 512-bit hashes is easily possible, but has to be statistically tested. The extension of the hash function to 128-bit math also works well: this increases its properties exponentially.
Passes all smhasher tests for a tiny size (157 byte on x64), but slow as every more secure hash. I would put it as alternative to people using siphash. Personally I see no use case for siphash at all. Slow, big, insecure, easily brute forcable.
PRVHASH can generate 32- to unlimited-bit hashes, yielding hashes of roughly equal quality independent of the chosen hash length. PRVHASH is based on 64-bit math. Hashes beyond 256-bits still require extensive testing, but, for example, any 32-bit element extracted from 512- or 2048-bit resulting hash is as collision resistant as just a 32-bit hash. The use of the function beyond 512-bit hashes is easily possible, but has to be statistically tested. The extension of the hash function to 128-bit math also works well: this increases its properties exponentially.
Passes all smhasher tests for a tiny size (157 byte on x64), but slow as every more secure hash. I would put it as alternative to people using siphash. Personally I see no use case for siphash at all. Slow, big, insecure, easily brute forcable.