University of Warwick researchers have been studying and comparing the supercomputing systems in China and the United States to provide an analysis that will benefit the battle plans of both sides in an escalating war between two competing technologies. The researchers compared the general-purpose graphics processing units (GPUs) used in China's 2.5 Petaflops Tianhe-1A with the BlueGene supercomputing designs used in the United States.
Led by Warwick professor Stephen Jarvis, the researchers used mathematical models, benchmarking, and simulations to determine the likely performance of the computing designs. "In our paper we show that BlueGenes can require many more processing elements than a GPU-based system to do the same work," Jarvis says. The researchers found that small GPU-based systems solved problems between three and seven times faster than traditional CPU-based designs.
"Given the crossroads at which supercomputing stands, and the national pride at stake in achieving exascale, this design battle will continue to be hotly contested," Jarvis says.
From University of Warwick
View Full Article
Abstracts Copyright © 2010 Information Inc., Bethesda, Maryland, USA