Today brings a new TOP500 list of the 500 fastest supercomputers in the world. And China still sits on top, with the number one and number two systems: Sunway TaihuLight, at a Linpack rating of 93 petaflops, and Tianhe-2, at 34 petaflops.
That said, two new systems are in the top 10, both of which run Intel Knights Landing Xeon Phi 7250 68-core processors: Cori, a new Cray XC40 at Berkeley Lab’s National Energy Research Scientific Computing Center (NERSC), lands at number five with a rating of 14 petaflops. Oakforest-PACS, a Fujitsu Primergy CX1640 M1 cluster, scored 13.6 petaflops; it resides in Japan’s Joint Center for Advanced High Performance Computing.
“In addition to matching each other in system count in the latest rankings, China and the US are running neck and neck in aggregate Linpack performance,” the TOP500 team said in a statement. The group said the US holds the narrowest of leads, with 33.9 percent of the total, and that China is second with 33.3 percent. The total performance of all 500 computers on the list adds up to 672 petaflops, 60 percent higher than it was a year ago; in June, it was 566 petaflops.
Other key data: Of the 500 systems now in the list, Intel chips power 462 of them. IBM Power chips account for 22, and AMD has just seven. Manycore accelerators are present in 96 of the systems, with 60 using Nvidia GPUs, 21 using Xeon Phi, one using AMD FirePro, one with PEZY, and three employing a combo of Nvidia and Intel Xeon Phi GPUs. Gigabit Ethernet interconnect is present in 206 systems, while InfiniBand technology accounts for 187. Intel’s relatively new Omni-Path went up from eight systems to 28 this time around.
Here’s the current list of the 10 fastest supercomputers in the world:
1. TaihuLight: Sunway MPP, SW26010; National Supercomputing Center in Wuxi, China; 10.6 million cores (93.01 petaflop/s).
2. Tianhe-2: TH-IVB-FEP Cluster; National Super Computer Center in Guangzhou, China; 3.12 million cores (33.86 petaflop/s).
3. Titan: A Cray XK7 system at the Department of Energy’s Oak Ridge National Laboratory (17.59 petaflop/s).
4. Sequoia: An IBM BlueGene/Q system located at the Department of Energy’s Lawrence Livermore National Lab in California, with 1.57 million cores.
5. Cori: A Cray XC40 at Berkeley Lab’s National Energy Research Scientific Computing Center (NERSC) (14 petaflop/s).
6. Oakforest-PACS: A Fujitsu Primergy CX1640 M1 cluster located at Japan’s Joint Center for Advanced High Performance Computing (13.6 petaflop/s).
7. K Computer: A SPARC64 system with 705k cores at RIKEN Advanced Institute for Computational Science in Japan (10.5 petaflop/s).
8. Piz Daint: Cray XC30 with 116k Xeon and Nvidia cores; located at the Swiss National Computing Centre in Switzerland; new Nvidia P100 GPUs boosted this one by 3.5 petaflops, so it’s still at number eight (9.8 petaflop/s).
9. Mira: IBM BlueGene/Q; DOE/SC/Argonne National Laboratory, US; inside are 786k custom IBM cores (8.6 petaflop/s).
10. Trinity: A Cray XC40 at DOE/NNSA/LANL/SNL in the US; it contains 301,056 Xeon E5-2698v3 cores (8.1 petaflop/s).
Top image credit: Jack Dongarra, Report on the Sunway TaihuLight System, June 2016