The Cryogenic Challenge : A "Real-Life" Parallel Application

The path integral problem is almost completely separable, meaning that each "bead" in the simulation could go to its own processor with a minimum of communication needed. Because of this situation, it is possible to achieve nearly perfect speedup. Specifically, on 16 processors, the problem ran 15 times faster than it did on one processor. Because of this, the entire research problem was reduced from a decade-long burden to a concise, one-year project.

I think this example shows clearly the overwhelming advantage that computational scientists now enjoy over their predecessors courtesy of parallel algorithms and platforms. With research projects finishing faster and faster, we are able to generate an appreciably greater volume of results. As it turns out, such an avalanche of numbers presents its own problems, namely finding ways to analyze the data. Next time around, we'll take a look at some specific ways in which advanced scientific visualization techniques are helping to ease the burden of this daunting task.

Go To Page: 1 2

Articles in this Topic    Discussions in this Topic