by Anton Shilov
07/01/2013 | 06:47 PM
Nvidia Corp. has announced that Nvidia Tesla GPU accelerators are powering the world's two most energy efficient supercomputers, based on the latest Green500 list published last week. Both systems are based on Nvidia Tesla accelerators based on Kepler family of graphics chips.
The winning system is 'Eurora' at Cineca, Italy's largest supercomputing center, in Casalecchio di Reno. Equipped with Nvidia Kepler architecture-based GPU accelerators, Eurora delivers 3210MFlops per watt, making it 2.6 times more energy efficient than the best system using Intel CPUs alone (at Météo France). It also greatly surpasses the most efficient Intel Xeon Phi accelerator-based system, ‘Beacon’, at the National Institute for Computational Sciences, at the University of Tennessee.
The number-two system on the June 2013 Green500 list is the 'Aurora Tigon' supercomputer at the Selex ES facilities Chieti, Italy.
These top two systems are powered by advanced Eurotech high-performance 'Aurora' servers, equipped with Nvidia Tesla K20 GPU accelerators.
More broadly, Nvidia Kepler GPU accelerator-based systems in the top 100 spots on the Green500 list are 50% percent more energy efficient on average than the latest Intel Xeon Phi-based systems.
Following last week's ISC 2013 student cluster challenge in which Nvidia GPU accelerator-based systems swept the top four spots for energy efficient supercomputing, Nvidia GPUs again demonstrated why they are most preferred accelerator solution among developers, system designers, and computational researchers.
"Raw performance is no longer the exclusive measure of the value and impact of supercomputers. All future systems will need to deliver higher performance with reduced power consumption. With GPU accelerators in the top two spots, the latest Green500 list demonstrates accelerators' ability to deliver unmatched levels of energy-efficient supercomputer performance for next-generation systems," said Sumit Gupta, general manager of the Tesla accelerated computing business unit at Nvidia.
By delivering thousands of small energy efficient cores operating in parallel, GPU accelerators are considerably more energy efficient than standard CPUs. These GPU cores are optimized to run the compute-intensive portions of applications, while general-purpose CPU cores, which process in serial, are highly inefficient when running compute-intensive applications.