The supercomputing arms race is heating up again between the United States and China, as China retakes the top spot in the 41st Top500 listing of the world's most powerful supercomputers with Tianhe-2, an updated system that was able to execute 33.86 petaflops, or 33.86 thousand trillion floating point operations per second.
Tianhe-2 churned out almost twice as many petaflops as the second computer on the list, Titan, a Cray XK7 system installed at the U.S. Department of Energy's (DOE) Oak Ridge National Laboratory. The top system in the previous Top500 competition, compiled in November, Titan clocked 17.59 petaflops on the Linpack benchmark for this round of the Top500.
The appearance of Tianhe-2 came as something of a surprise, even as word about its immense capabilities leaked out on the Internet earlier this month. The system was not expected to be operational for another two years.
China's National University of Defense Technology built Tianhe-2, or Milky Way-2, with 16,000 nodes. Each node runs two Intel Xeon IvyBridge processors and three Xeon Phi processors, for a combined total of 3.12 million computing cores. The system, located at the National Supercomputer Center in Guangzho, China, will be fully operational by the end of the year.
The last time China had topped the list -- which is often capped by a U.S. Energy Department machine -- was with Tianhe-1A, which occupied the top position in November 2010.
The U.S. remains the dominant supercomputer leader, with 253 of the 500 systems on the list. With 65 systems on the list, China ranks second, ahead of Japan, the U.K., France and Germany.
Besides challenging the U.S. dominance of the Top500, The Tianhe-2 system is also notable for its use of technologies developed in China.
"Most of the features of the system were developed in China, and they are only using Intel for the main compute part. The interconnect, operating system, front-end processors and software are mainly Chinese," said Top500 editor Jack Dongarra in a statement. Dongarra visited the site of the Tianhe-2 system in May.
Coming in the third spot is the DOE Lawrence Livermore National Laboratory's Sequoia, an IBM BlueGene/Q system that achieved 17.17 petaflops on the Linpack benchmark using 1.5 million cores. Overall, four IBM BlueGene systems made the top 10 list.
Other interesting tidbits about this edition of the Top500: Twenty-six petaflop systems are now on the list, up from 23 six months ago. Eighty-eight percent of the systems use processors of six or more cores, and 80% of the systems use Intel processors. The total combined performance of all 500 systems is 223 petaflops, up from 162 petaflops six months ago.
This edition of the Top500, a twice-yearly ranking of the world's fastest publicly known supercomputers, was announced during the 2013 International Supercomputing Conference in Leipzig, Germany.