Thanks for the link. Like the article mentioned, half the battle is figuring out how to divide up the problem so it can be processed efficiently. Once you learn how to do that, you can solve all sorts of problems.

This clustering of inexpensive computers is getting supercomputer performance to small groups but it's also killing the supercomputer manufacturers. I think that was one of the reasons why Semour Cray never managed to sell one of his Cray-3 machines. The Cray-3 was very fast and used gallium-arsenide technology but was too expensive. He tried to salvage the company with the a new, less expensive design called the Cray-4 but died before making that happen. He died here in Colorado Springs following an auto accident.