So far I can find a wide ranging upper and lower bound for my lowest co prime in a 2 co prime composite then I can find a small bunch of smaller bounding's within. The Question I have is can I then end up with a fractal like program wher first it analyses the 2 co prime composite then it does the same analysis in the smaller bounds until I can go no future with my number and my prime will be in one of a far larger range of tiny bounding spots upon which I then use a Gambling function which will in the end find my co prime but that with the right timing and parallelization gives me better odds then a sieve'ing mechanism on top of the efficiency benefits of only having to look in small crevasses.
This should save me much time and memory considering what I already have achieved when searching for a semi prime/co prime.
Another tool which might up my odds if tuned right could be the what I call MMod or multimod.
So MMod or MultiMod takes and orders from largest to smallest a range of mods which it adds up and as the main Moding and divides into subsections equal to and between the ranges. This technique offers the ability the ability of of having a two tier Mod function.
This from Idea from even very early testing is already displaying interesting results which could be very useful for gambling fast and with good odds of finding the desired number quickly when use in a Gambling test.