I am facinated by the power of a grid of computers to solve numerical problems faster than a single computer could do.
As a result, I picked a little example to compare the possibilities of some grid computing solutions I stumbled about in the last few months. Since I am used to develop in Java, the solutions are all Java based.
A partitionable problem
The first thing to consider is to find a partitionable problem the grid could work on. One typical example for such a problem are fractals, especially the classical one, the one of Mandelbrot (Mandelbrot Wiki). A lot of “pixels” have to be computed and each pixel can be processed individually.