It's easy to press a few buttons to come up with the strange looking result of 9366. With n=57 this is accurate to at least 6 decimal places.
It's hard to derive this result analytically. In fact, I'd assume there is a formula for any a>0 and b>1 for the sum of j^a/2^b.

Posted by Jer
on 20110422 14:40:04 