Okay, I need someone interested in speed benchmarks to help me in comparing the latest GLIBC memmem() against my Railgun_Trolldom.
I optimized 'Trolldom' a lot, the only thing that chokes it is the speed of RAM fetching. I have only laptops with slowish 666MHz RAM, so no real picture of how 100TB of data are parsed.
Since I have no powerful toys to play with, as Skylake and AMD Zen, please run the benchmark for C community's sake, I will publish the results on: http://www.codeproject.com/Articles/250566/Fastest-strstr-like-function-in-C
The benchmark is downloadable at my Internet drive and my site: https://1drv.ms/u/s!AmWWFXGMzDmEgl3izdwhSnIq0Lv5
Just go to console prompt and run 'GNU_vs_Railgun.bat', it takes a lot of time but it gives memmem() behavior in real-life application.