Speeding up bottleneck

[Keith Goodman]

I'm playing around with SSE3, AVX, and OpenMP to see which techniques might help bottleneck.

​

I've seen some encouraging initial results on linux

​: https://github.com/kwgoodman/some_sums

I'm now wondering if I can get the some_sums repo to compile on Windows and OSX. So far I haven't had any luck on travis or appveyor.

If you're interesting in having a look, the github issue is here: https://github.com/kwgoodman/some_sums/issues