Cellular Automata Modeling the "Only Up" Collatz Problem

This implements the no-descent collatz function. At each iteration, add on to n the product of its even prime factors, then apply 3n+1. The collatz conjecture equivalent on this frame is that each number will eventually become a power of 2. In the automata above, this is a single on bit in the row.

Example of product of even prime factors: 14 is 2, 24 is 8, 8 is 8, 7 is 1. Its just the number of 2s that divide it.

Notice that flipping a bit in the seed (top row), every other bit in that column flips, and therefore every bit to the left and down from it also flips. This is what makes this automata more workable than the traditional collatz automata, where causal relations are not so easily seen.

This is the program that convinced me the collatz conjecture is probably false.