If you can immediately come up with an algorithm for a well-studied computer science problem, then it's very likely the approach isn't going to work out. Notab

If you can immediately come up with an algorithm for a well-studied computer science problem, then it's very likely the approach isn't going to work out. Notably in your case you've just converted high-precision floating point into even-higher-precision integer math. So the problem here that's inherent, which it that you might have to look at a very large number of bits of precision to find the differences, hasn't been sidestepped.