SCIENCE

The November 2010 list of the "Top500" fastest supercomputers in the world was released and it revealed that NVIDIA Tesla GPUs are now powering three of the top five systems.

Tesla GPUs were featured in the number one, three and four slots with the recently announced Tianhe-1A system taking the top spot with a performance record of 2.507 petaflops.

The top three GPU supercomputers deliver more performance than the rest of the Top 10 systems combined. The most notable new entry to the TOP500 is Tsubame 2.0, the new supercomputer from Tokyo Institute of Technology. The system delivers petaflop-class performance while remaining extremely efficient, consuming just 1.340 megawatts, dramatically less power than any other system on the top five.

"Tsubame 2.0 is an impressive achievement, balancing performance and power to deliver the most energy efficient petaflop-class supercomputer ever built," said Bill Dally, chief scientist at NVIDIA. "The path to exascale computing will be forged by groundbreaking systems like Tsubame 2.0."

GPUs have quickly become the enabling technology behind the world's top supercomputers. They contain hundreds of parallel processor cores capable of dividing up large computational workloads and processing them simultaneously, significantly increasing system performance. Heterogeneous systems, built with GPUs and CPUs, require less space and consume less power, making supercomputing more affordable and more accessible than ever before.

Dally is the Plenary speaker at this week's SC'10 supercomputing conference in New Orleans and will present on Wednesday, Nov. 17 on the subject of "GPU Computing: To Exascale and Beyond."