[https://lemire.me/blog/2018/08/22/avoid-lexicographical-comparisons-when-testing-for-string-equality](https://lemire.me/blog/2018/08/22/avoid-lexicographical-comparisons-when-testing-for-string-equality)/
Output of running his code is
$ ./memcmpfun
mass_comparison(bigarray1, bigarray2,comparisonsA, N) :
8.296 cycles per operation (best) 2.411 bytes per cycle (best) 43.689
cycles per operation (avg)
mass_comparison_fast(bigarray1, bigarray2,comparisonsB, N) :
3.778 cycles per operation (best) 5.294 bytes per cycle (best) 4.356
cycles per operation (avg)
mass_comparison_faststruct(bigarray1, bigarray2,comparisonsB, N) :
4.667 cycles per operation (best) 4.286 bytes per cycle (best) 6.015
cycles per operation (avg)
mass_comparison_bcmp(bigarray1, bigarray2,comparisonsB, N) :
3.333 cycles per operation (best) 6.000 bytes per cycle (best) 3.793
cycles per operation (avg)
mass_comparison_hash(bigarray1, bigarray2,comparisonsB, N) :
8.000 cycles per operation (best) 2.500 bytes per cycle (best) 10.104
cycles per operation (avg)
Run
I have no idea how he gets the exact cycle count -- but its all contained in
200 lines of code in file benchmark.h
