these implement just enough blas-like routines to implement a solver over a finite field of octets.
gradually optimizing for different archs. (eg. SSE, AVX, NEON)
references: https://www.ssrc.ucsc.edu/Papers/plank-fast13.pdf
unibg-seclab/oblas
small c99 blas inspired routines for finite field algebra over gf(256)
Objective-CMIT