Aleksey Vaneev

> Would you be willing to create a pull request? Sorry, I'm not very experienced with collaborative features of GitHub. Please simply copy the file to your repository, its license...

@lemire There's a new finding about PRVHASH, maybe worth testing as well. "Fused PRNG" arrangement. It shows me 0.41 cycles/byte which is exceptional for a non-SIMD PRNG. It works fine...

I would run the test for 10000000 iterations, but this requires to keep at least Len&RandomCount unique over the test. So, it's just a loop with inner loop: for( int...

The buffer can be pre-allocated and only zeroed for the required Len.

@rurban here is the implementation of `sparse long` buffer filling: ``` int l=...; int c= ...; memset( &SparseBuf[ 0 ], 0, l ); int i; for( i = 0; i...

Isn't SHA2 seed-less? How did you manage to add seed to it?

The same question applies to SHA3 actually. Is seeding documented?

This is useful information, but how to evaluate it is a complex question. E.g. I had two variants of hashing code that showed 0.71% average, but while one gave 1.5...

Hello! What are the use-cases for the streamed version? At this point I'm not very keen on producing a streamed version. I'm also thinking that concatenation of strings in a...

Okay, I'll think about a better way to implement that. At the moment you may take a look at PRVHASH64S: https://github.com/avaneev/prvhash It offers about 8.4 GB/s streamed hashing throughput and...