Home > Plugged In > Archives > 2008 > November > 17 > Entry
IBM and Cray take fastest supercomputer spots
Roadrunner is still the fastest animal in the supercomputing world, but Jaguar is nipping at its heels.
IBM’s Roadrunner system kept its hold on the No. 1 spot on the latest Top500.org list of the world’s fastest supercomputers. Cray’s XT5 Jaguar system, however, has become the first to join it beyond the petaflop mark.
The list is scheduled to be released today at the SC08 supercomputing conference, which is being held through Friday at the Austin Convention Center. The event is expected to draw 9,000 people from around the world.
Austin’s own supercomputer - the “Ranger” system at the University of Texas’ advanced computing center - dropped two places on the list, but remained in the top 10 at No. 6.
With the aid of a few enhancements added since June, the Roadrunner system at the Los Alamos National Laboratory pushed its performance on a benchmark test to 1.105 petaflops.
Normally, we’d try to come up with a clever way to portray the magnitude of that number. In this case, just typing it out seems impressive enough - 1,105,000,000,000,000 floating point operations per second.
(Generally speaking, a desktop PC might pump through roughly 10,000,000,000 “flops.” What a puny little number that is.)
The supercomputer runs all those calculations through a combination of IBM Cell and AMD Opteron processors. The bulk of the design work for both processors occurred at those companies’ offices in Austin.
Roadrunner became the first supercomputer to surpass the petaflop barrier earlier this year. Cray now has joined it.
The Jaguar supercomputer, located at the Oak Ridge National Laboratory, clocked in at 1.059 petaflops. Again, for effect: That’s 1,059,000,000,000,000 floating-point operations per second.
Cray and IBM dominate the top 10, taking seven of those spots (Cray 4, IBM 3).
Author= Dan Zehr Author_email= dzehr@statesman.com
IBM, Cray, AMD, supercomputer, Top500, Los Alamos National Laboratory, Oak Ridge National Laboratory, University of Texas, SC08
Permalink | Comments (0) | Post your comment | Categories: Texas
Comments