Tuesday, June 24, 2003 08:37 AM
I have built a simple supercomputer that runs via RMI. I have used this to calculate and display Mandelbrot diagrams. This is straightforward as I can break the problem down into a set of atomic subproblems. I now wish to go further and model Conway's Game of Life. This requires that sub problems be they single cells or blocks of cells exchange data. Any advice upon what strategy / architecture that should be used so that I can minimise the cost of RMI? Any pointers to websites / groups etc that deal with this issue?