3 ms·

Of course, the next step is obvious - work out why the compiler didn't do a four way avx unroll, and then submit a bug fix to clang to make it do that. That way

Of course, the next step is obvious - work out why the compiler didn't do a four way avx unroll, and then submit a bug fix to clang to make it do that. That way all of your future code benefits from your single micro-optimization. It's also possible that you find out that if you enable --generate-for-haswell or some other arcane compiler flag, it'll do it for you.

pbsd 12y ago
All the author had to to was to add '-march=native' or '-march=core-avx2' to the compiler command line: http://goo.gl/H4f62I http://goo.gl/H4f62I
- jevinskie 12y ago
  Clang 3.7.0 (experimental) + -march=skylake gives you AVX512. zmm all the way, baby! 256 bytes processed in the inner loop!
  - pantalaimon 12y ago
    But what gives me a Skylake CPU?
    - duskwuff 12y ago
      A time machine, or a job working at Intel? :)