Recent content by Ulam

If you want be happy, you can arrest the algorithm at \sqrt{n}.

In this case the problem is a P problem in information theory, in other words the time to calculate p(n,l,k) is Polynomial, so it will spend more time to calculate p(millions,thousands,thousands). Only a formula will give you a gift, but it is not discovered yet, and maybe it's likely that it...

The question was if there is a "nice" formula, not a "nice" algorithm. The formula, for now, isn't discovered, maybe it doesn't exists, but the algorithm to calculate p(n,l,k) is easy to make.

Nobody knows, but it seems unlikely that a formula exists.