China Takes Top Spot... Again

China once again will be the home of the fastest supercomputer on the planet. The Tianhe-2 (Milkyway-2) system, housed in a facility in Changsha, has exceeded 30 petaFLOPS worth of computational capacity in a recent benchmark test and has a theoretical peak performance of 54.9 petaFLOPS. The Tianhe-2 has 32,000 Intel Xeon Ivy Bridge processors and 48,000 Xeon Phi co-processors. The Xeon Phi co-processor is the first chip based on Intelís Many Integrated Core (MIC) architecture.

The Tianhe-2 takes the top spot away from Oak Ridge National Laboratoryís Titan system. Titan is a Cray XK7 system comprised of 299,008 AMD Opteron cores and 18,688 Nvidia Tesla K20 accelerators. Titanís current Linpack benchmark is just under 18 petaFLOPS with a theoretical peak of 27 petaFLOPS. This makes the Tianhe-2 system about twice as fast as the Titan system.

US researchers warned Congress last month that the US is at risk of falling behind in HPC development unless it commits hundreds of millions of dollars to exascale research. Unfortunately the ongoing budget disputes are leading to a reduction in R&D spending in the US while the Chinese are increasing theirs.

Jack Dongarra, a professor of computer science at the University of Tennessee and one of the academic leaders of the Top 500 list, stated that the next large supercomputer acquisition for the US Department of Energy will not be until 2015. Looks like the US will need to get used to not being on the top for a while.
ClusterChimps © 2009