Tuesday, July 19, 2005

My biggest simulation yet

I just submitted the most important simulation I've prepared so far. It will run parallel in 32 dual Xeon processors at USC's Hyper Performance Computer Cluster. It will need about 15 hours to complete, which is a ridiculus number for a 1D simulation. The reason it takes so long is that I need hyper-fine resolution on my parameters, since all the details have to be captured and also it has to be resilient to numerical noise (the situation improves the higher the resolution since nearby points are closer together hence the transition between them is smoother).

If my calculations are correct, the simulation will be done by the time I get back from Vegas. Viva!