See https://blake2.net/ for information about BLAKE2

This is an implementation of BLAKE2 for PowerPC AltiVec.
It achieves 6.1 cycles/byte** on a puny 1.2 GHz PPC G4 7447A,
which is faster than SHA-1. (Using Linux 3.2, gcc 4.7).

Of course I'll be happy if there are any ways to improve upon my altivec core.

BLAKE2s features implemented are: sequential, salted, keyed, arbitrary digest
size.


to build:

    make blake2s-altivec

run testsuite:

    ./blake2s-altivec

checksum:

    ./blake2s-altivec FILE

benchmark (Linux):

    perf stat -- ./blake2s-altivec --bench

**From an in-memory benchmark. 6.1 cycles/byte is equivalent to 185 MB/s.
Hashing an actual file from (cached) disk I/O results in 142 MB/s.