Merge pull request #22 from Jille/not

Implement bitwise Not

Merge pull request #21 from Jille/vzeroall

Call VZEROALL for a 21% benchmark improvement

Merge pull request #20 from Jille/avx1

Add AVX (not AVX2) implementations

Merge pull request #16 from Jille/new-benchmarks

Update benchmarks in README now that our purego memset is optimized

Merge pull request #4 from Jille/fixes

Make the test actually validate the code

Merge pull request #1 from Jille/buildtag

Remove unneeded + broken build tag