I think this is the most important aspect of this paper. Throwing more computing power at the problem increases performance significantly. It is possible that our algorithms are adequate but our hardware is not.