For the second time in a row there is no system from the USA under the TOP3. #1 and #2 are installed in China, a system in Switzerland now at #3, and a new system in Japan is #4 pushing the top US system to #5 since last last June.
Sunway TaihuLight, a system developed by China’s National Research Center of Parallel Computer Engineering & Technology (NRCPC) and installed at the National Supercomputing Center in Wuxi, which is in China's Jiangsu province maintains the lead as the No. 1 system with 93 petaflop/s (Pflop/s) for the fourth time.
Tianhe-2 (Milky Way-2), a system developed by China’s National University of Defense Technology (NUDT) and deployed at the National Supercomputer Center in Guangzho, China is now the No. 2 system with 33.86 Pflop/s. Tianhe-2 was the No.1 system in the TOP500 list for the past 3 years (6 lists)
The No. 3 is the Piz Daint, a Cray XC50 system installed at the Swiss National Supercomputing Centre (CSCS) in Lugano, Switzerland and the most powerful system in Europe. The system has a total of 361,760 cores.
The upgraded Gyoukou a ZettaScaler-2.2 HPC system using PEZY-SC2 to accelerate its computations achieved now 19.1 Pflop/s and is the new No. 4 system. It utilizes 20,000 conventional Intel Xeon cores together with 19,860,000 PEZY-SC2 cores operating at 500 MHz. This is the highest level of concurrency seen in the TOP500. The system is installed at the Japan Agency for Marine-Earth Science and Technology (JAMEST), which was the home for the Earth Simulator.
Titan, a Cray XK7 system installed at the Department of Energy’s (DOE) Oak Ridge National Laboratory, and the largest system in the USA. is now the No.5 system. It achieved 17.59 Pflop/s using 261,632 of its NVIDIA K20x accelerator cores.
Sequoia, an IBM BlueGene/Q system installed at DOE’s Lawrence Livermore National Laboratory, is the No. 6 system. It was first delivered in 2011 and has achieved 17.17 Pflop/s using 1,572,864 cores.
Trinity a Cray XC40 system operated by Los Alamos National Laboratory and Sandia National Laboratories and located at Los Alamos was upgraded to 940,800 cores and achieved 14.1 Pflop/s, which puts it at the No. 7 position.
Cori, a Cray XC40 supercomputer comprised of 1,630 Intel Xeon "Haswell" processor nodes, 9,300 Intel Xeon Phi 7250 ("Knight's Landing") nodes entered the TOP500 in November 2016 and is now at No. 8 with 14.01 Pflops/s using 622,336 cores.
Oakforest-PACS, a Fujitsu PRIMERGY CX1640 M1 installed at Joint Center for Advanced High Performance Computing in Japan is powered by Intel Xeon Phi 7250 nodes and Intel Omni-Path interconnect technology is at No. 9 with 13.55 PFlop/s using using 558,144 cores.
Fujitsu’s K computer installed at the RIKEN Advanced Institute for Computational Science (AICS) in Kobe, Japan, is now the No. 10 system with 10.51 Pflop/s using 705,024 SPARC64 processing cores.
Highlights from the Overall List
The number of system in the USA decreased further to a new record low of 143 from 169 six month ago.
The number of systems installed in China increased to a new record high of 202, compared to 160 on the last list. China now clearly show a substantially larger number of installations then the USA.
China now is also pulling ahead of the USA in the performance category with China holding 35.4% of the overall installed performance while the USA is second with 29.6% of the overall installed performance.
There are 181 systems with performance greater than a Pflop/s on the list, up from 138 six months ago.
In the Top 10, the No. 2 system, Tianhe-2, the No. 7 Trinity, the No. 8 Cori, and the No. 9 Oakforest-PACS uses Intel Xeon Phi processors to speed up their computational rate. The No. 3 system Piz Daint and the No. 5 system Titan are using NVIDIA GPUs to accelerate computation.
A total of 102 systems on the list are using accelerator/co-processor technology, up from 91 on the June 2017 list. 86 of these use NVIDIA chips, 12 systems with Intel Xeon Phi technology (as Co-Processors), and 5 are using PEZY technology. Two systems use a combination of Nvidia and Intel Xeon Phi accelerators/co-processors. An additional 14 Systems now use Xeon Phi as the main processing unit.
The average number of accelerator cores for these 102 systems jumped now to 294,000 cores/system driven by the large number of cores in PEZY accelerated systems.
Intel continues to provide the processors for the largest share (94.2 percent) of TOP500 systems.
Ninety-three (95.8) percent of the systems use main processors with eight or more cores, sixty-eight (76.0) percent use twelve or more cores, and twenty-seven (38.2) percent 16 or more cores.
We have incorporating the HPCG benchmark results into the Top500 list to provide a more balanced look at performance.
The fastest system on the HPCG benchmark remains Fujitsu’s K computer which is ranked #10 in the overall TOP500. It is again followed closely by Tianhe-2 which is also No. 2 on the TOP500.
General highlights from the TOP500 since the 49th edition
The entry level to the list moved up to the 548 Tflop/s mark on the Linpack benchmark, compared to 432 Tflop/s six months ago.
The last system on the newest list was listed at position 370 in the previous TOP500. This turnover is in line with what has been seen during the last four years, but much lower than previous levels.
Total combined performance of all 500 systems has grown to 845 Pflop/s, compared to 749 Pflop/s six months ago and 672 Pflop/s one year ago. This increase in installed performance is well below the previous long-term trend we had seen until 2013.
The entry point for the TOP100 increased in six months to 1.28 Pflop/s, up from 1.21 Pflop/s.
The average concurrency level in the TOP500 is 138,000 cores per system, up from 96,160 six months ago and 87,990 one year ago.
A total of 471 systems (94.2 percent) are now using Intel processors, slightly up from 92.8 percent six months ago.
The share of IBM Power processors is now at 14 systems, down from 21 systems six months ago.
Gigabit Ethernet is now at 228 systems (unchanged), in large part thanks to 204 systems now using 10G interfaces. InfiniBand technology is now found on 163 systems, down from 178 systems, and is the second most-used internal system interconnect technology.
Intel Omni-Path technology is now at 35 systems down from 38 system six month ago.
HPE has the lead in systems and now has 122 systems (24.4 percent). This count for HPE includes several systems originally installed by former SGI. HPE had 144 systems six months ago.
Lenovo follows with 81 systems down from 88 systems.
Inspur rose further in the ranks and has now 56 systems (11.2 percent) up from only 20 six month ago.
Cray now has 53 systems, down from 57 systems six month ago.
Sugon features 51 systems in the list up from 44.
IBM follows with only 19 systems remaining under their label. These are mostly BlueGene/Q systems and the average age of IBM systems on the list is now 5 years.
Cray continues to be the clear leader in the TOP500 list in performance and has a considerable lead with a 19.5 percent share of installed total performance (down from 21.4 percent).
Due to SGI based systems, HPE is second with 15.2 percent, down from 16.7 percent six months ago.
Thanks to the Sunway TaihuLight system, NRCPC retains the third spot with 11.1 percent of the total performance (down from 12.5 percent).
Lenovo is fourth with 9.1 percent of performance.
Inspur rose to fifth with 6.3 percent.
IBM is now in the sixth spot with 6.1 percent share.
And Sugon rose to 5.2 percent.
China is now the leading consumer of HPC systems with 202 systems (up from 160) ahead of the USA at 143 systems (down from 169). The European share (93 systems, down from 106 in the last list) is noticeable lower than the Asian share of 251 systems, up from 210 six month ago.
Dominant countries in Asia are China with 202 systems and Japan with 35 systems (up from 33).
In Europe, Germany is the leader with 21 systems followed by France with 18 and the UK with 15 systems.
The data collection and curation of the Green500 project has been integrated with the TOP500 project. This allows submissions of all data through a single webpage at http://top500.org/submit
The top 3 positions in the Green500 are all taken by newly installed systems in Japan. All of these systems are based on the ZettaScaler-2.2 architecture.
The fourth position is held by the NVIDIA DGX-1 system at NVIDIA.
The fifth position is held by the new No. 4 system of the TOP500 Gyoukou also a ZettaScaler-2.2 system.
It is followed by Tsubame 3.0, a modified HPE ICE XA System at the GSIC Center, Tokyo Institute of Technology, Japan. This system was the No. 1 on the Green500 six month ago.
The most energy-efficient system and No. 1 on the Green500 is the new Shoubu system B, a ZettaScaler-2.2 system at the Advanced Center for Computing and Communication, RIKEN, Japan. It achieved 17.0 GFlops/Watt power-efficiency during its 788 Tflop/s Linpack performance run. It is listed on position 258 in the TOP500.
No. 2 in the Green500 is the Suiren2 system at the High Energy Accelerator Research Organization/KEK, Japan. This ZettaScaler-2.2 system achieved 16.8 GFlops/Watt. It is listed on position 306 in the TOP500.
No. 3 in the Green500 is the Sakura system installed the manufacturer of the system PEZY Computing K.K., Japan. It achieved 16.7 GFlops/Watt. It is listed on position 275 in the TOP500.
The first system outside of Japan is the No. 4 the DGX SaturnV Volta system, a Nvidia System installed at NVIDIA, USA. It achieve 15.1 GFlops/Watt power efficiency. It is on position 149 in the TOP500.
The No. 5 system is Gyoukou, another ZettaScaler-2.2 system installed at the Japan Agency for Marine-Earth Science and Technology, Japan. It achieved 14.2 GFlops/Watt power efficiency. It is on position 4 in the TOP500.
The Top500 list now includes the High-Performance Conjugate Gradient (HPCG) Benchmark results.
The top 5 systems on the HPCG list are: the K computer (#10 on the Top500), Tianhe-2 (#2 on the Top500), Trinity (#5 on the Top500), Piz Daint (#3 on the Top500), and the Sunway TaihuLight (#1 on the Top500).
The International Space Station computer is listed in the HPCG benchmark. Making it the “highest” computer on the HPCG list.
About the TOP500 List
The first version of what became today’s TOP500 list started as an exercise for a small conference in Germany in June 1993. Out of curiosity, the authors decided to revisit the list in November 1993 to see how things had changed. About that time they realized they might be onto something and decided to continue compiling the list, which is now a much-anticipated, much-watched and much-debated twice-yearly event.
The TOP500 list is compiled by Erich Strohmaier and Horst Simon of Lawrence Berkeley National Laboratory; Jack Dongarra of the University of Tennessee, Knoxville; and Martin Meuer of ISC Group, Germany.