China’s Tianhe-2 supercomputer once again sits atop the list of the world’s fastest systems, a position it now has held for more than two years.
The latest Top500 list of the world’s fastest supercomputers was released July 13 at the International Supercomputing Conference (ISC) in Germany, and Tianhe-2—also known as Milky Way-2—kept its top spot with a performance of 33.86 petaflops (quadrillions of calculations per second), about twice that of Titan, an XK7 system from Cray that came in second with a performance of 17.59 petaflops.
It was once again fairly static at the top of the Top500 list, with only one new entrant in the top 10—the Shaheen II installed at the King Abdullah University of Science and Technology in Saudi Arabia. The Shaheen II is an XC40 system from Cray that came in at No. 7 with a performance of 5.536 petaflops. It is the first system from the Middle East to break into the top 10.
Overall, the list that is all about speed highlighted some key areas that seem to be slowing, according to organizers. Outside of Shaheen II, the other systems in the top 10 were installed in either 2011 or 2012, continuing a slowing trend in turnover among the top systems that began in 2008, they said.
In addition, the total combined performance of all 500 systems on the twice-yearly list grew to 361 petaflops, a jump from the 309 petaflops on the list from November 2014 and the 274 petaflops from the June 2014 list. “This increase in installed performance also exhibits a noticeable slowdown in growth compared to the previous long-term trend,” the organizers said in a statement.
The United States continues to be the country with the most systems on the Top500 list, with 233 supercomputers, an increase over the 231 from November 2014. However, the overall trend right now is downward—the country had 265 systems on the list in November 2013, and list organizers said it is nearing its historical low of 226 in 2002.
While keeping hold of the top spot, China also is seeing its number on the Top500 list fall, from 61 in November 2014 to 37 in the latest list. Both Europe and Japan gained in the number of systems they have one the list.
The trend toward accelerators also is growing, according to organizers. Eighty-eight systems are using accelerators from Intel (x86 co-processors) or Nvidia or Advanced Micro Devices (GPUs) to help boost performance while keeping a lid on power consumption. That is an increase over the 75 on the November 2014 list that used accelerators.
Both the Tianhe-2 system and the eighth-fastest supercomputer—the Stampede system at the Texas Advanced Computing Center at the University of Texas—use Intel’s Xeon Phi co-processors. Titan and the sixth-fastest system—Piz Daint, at the Swiss National Supercomputing Center—use Tesla K20x GPU accelerators from Nvidia. Stampede is built using Dell’s PowerEdge C8220 servers while Piz Daint is based on Cray’s XC30 systems.
Overall, 52 systems use Nvidia GPU accelerators, while another four use AMD’s Radeon chips. Thirty-three leverage Intel’s Xeon Phi co-processors, while four use a combination of Nvidia GPUs and Xeon Phi chips.
Sixty-eight systems have a performance greater than 1 petaflop—an increase over the 50 on the November 2014 list—and 97 percent of the systems use chips with six or more cores. Almost 88 percent use chips with eight or more cores.
Hewlett-Packard continues to have the highest number of systems on the list, with 178—or 35.6 percent of the list. In November 2014, HP had 179 systems. IBM has 111 systems (22.2 percent), a drop from the 153 on the last list. Cray came in third, with 71 systems, or 14.2 percent of the list.