We will use a Monte Carlo style estimation of pi. The idea is to randomly sample points in a 1x1 square. If that point lands in the a circumscribed circle then it is is a hit. If we throw enough darts we might eventually get a decent approximation of pi. This is not the easiest way to calculate pi, but it is a good demo for using distributed computation.
The point of this demo is to use distributed computers to do a calculation. Every browser with this page open participates in a distributed cluster used to calculate pi, albeit the hard way.