Implementations of BLAKE2sp and BLAKE2bp

[pedrocr]

Apparently BLAKE2 includes parallel versions of the two hash functions to take advantage of multi-core:

https://github.com/BLAKE2/BLAKE2/blob/master/sse/blake2sp.c

It would be nice to have these to be able to take advantage of more cores when hashing large files.

[Ralith]

These would also be useful as SIMD implementations, enabling dramatic speedups on common hardware without any extra threads at all.

[tomtau]

btw, here's a bounty on this issue: https://www.bountysource.com/issues/46854589-implementations-of-blake2sp-and-blake2bp

[pczarn]

I started working on the issue.

I am currently making a script to convert test data from the BLAKE2 repo to blobs in the hashes repo.

[tarcieri]

FYI, I just rebased #228 which completely replaces the current implementation with blake2-simd

[pczarn]

@tarcieri Okay, I can see that blake2-simd includes both parallel variants.

[newpavlov]

Closing in favor of #1.