I seem to remember implementations like this already being in formation.
Jay Freeman (aka saurik) is working on one called Orchid with a handful of other big names IIRC. I think I read about it here on YC.
From a cryptography standpoint, could this hint at attack vectors for things like discrete log problems?
I've only learned of the math behind that myself recently, not sure what implications having a "topographical map of the primes" could have, especially if the pattern is relf-repeating regardless of the size of the primes.