Sign In

Communications of the ACM

ACM TechNews

Speed Within Reach at Affordable Prices


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook

Supercomputers have reached new power and performance benchmarks while also increasing their availability to businesses in the past year. In June, IBM's Roadrunner machine at the Los Alamos National Laboratory reached the petaflop level by processing more than 1 million billion calculations per second, while Cray's Jaguar supercomputer at Oak Ridge National Laboratory performed the same feat in November. Dave Turek of IBM's Deep Computing group estimates that processing power increases by a factor of 1,000 once every decade or so, which has subsequently supported more refined analysis in terms of output. "We're now getting to the point where a supercomputer isn't just used to model the behavior of a single component in a jet engine, for example, but whole collections of components, working together in an integrated system," Turek says. "Or we can model different processes on a much smaller or larger scale, from the effect of a new drug on a disease at a molecular level, to the impact of human behavior on climate change over hundreds of years." The boundary between traditional supercomputing and high-performance computing (HPC) has blurred since parallel processing became available to commercial organizations in the early 1990s, and IDC estimates that sales of HPC systems costing more than $500,000 climbed 24 percent last year. But low-end systems that cost less than $250,000 are emerging as the primary growth factor as the technology becomes increasingly important to industry, government, and academic research. Modern supercomputers are using off-the-shelf components and costs have fallen, making the systems more affordable for smaller organizations.

From Financial Times
View Full Article


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account