This is the 55th edition of the TOP500.
After a few editions with very little new entries in the Top10 we have a new #1 and a total of 4 new systems:
Supercomputer Fugaku, a system based on Fujitsu’s custom ARM A64FX processor is the new #1. It is installed at the RIKEN Center for Computational Science (R-CCS) in Kobe, Japan, the location of the former K-Computer. It was developed by Fujitsu in close collaboration with Riken and uses Fujitsu’s Tofu D interconnect to transfer data between nodes.
It achieved a spectacular 416 Pflop/s on the HPL benchmark easily exceeding the old #1 Summit by 2.8x. In single or half precision (16-bit), which are often used in machine learning and AI applications, it’s peak performance is actually above 1,000 PFlop/s (= 1 Exaflop/s) and because of this, it is often introduced as the first ‘Exascale’ supercomputer.
HPC5 at #6 is the second new system. It is a PowerEdge system build by Dell installed by the Italian company Eni S.p.A. and is now the most powerful system in Europe. It achieves its performance of 35.5 Pflop/s by using NVIDIA Tesla V100 as accelerators and Mellanox HDR Infiniband as network.
Selene at #7 is an NVIDIA DGX A100 SuperPOD installed in house at NVIDIA in the USA. Naturally it uses the new NVIDIA Ampere A100 for acceleration and a Mellanox HDR Infiniband as network as well to achieve it’s 27.6 Pflop/s.
Marconi-100 at #9 is the second Italian system in the Top10. It is an IBM Power System AC922 based on IBM POWER9 processors, Nvidia Volta V100 accelerators, and a Dual-rail Mellanox EDR Infiniband. The system is installed at the Italian research center CINECA and achieved 21.6 Pflop/s.
Supercomputer Fugaku, a system based on Fujitsu’s custom ARM A64FX processor is the new #1. It is installed at the RIKEN Center for Computational Science (R-CCS) in Kobe, Japan, the location of the former K-Computer. It was developed by Fujitsu in close collaboration with Riken and uses Fujitsu’s Tofu D interconnect to transfer data between nodes.
Summit and Sierra slipped to the #2 and #3 spots in the USA
Summit, an IBM-built system at the Oak Ridge National Laboratory (ORNL) in Tennessee, USA, remains the fastest system in the U.S. now at the #2 spot worldwide of 148.8 Pflop/s on the HPL benchmark, which is used to rank the TOP500 list. Summit has 4,356 nodes, each one housing two Power9 CPUs with 22 cores each and six NVIDIA Tesla V100 GPUs each with 80 streaming multiprocessors (SM). The nodes are linked together with a Mellanox dual-rail EDR InfiniBand network.
Sierra, a system at the Lawrence Livermore National Laboratory, CA, USA is now at #3. It’s architecture is very similar to the new #2 systems Summit. It is built with 4,320 nodes with two Power9 CPUs and four NVIDIA Tesla V100 GPUs. Sierra achieved 94.6 Pflop/s.
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 is now listed at the #4 position with 93 Pflop/s.
Tianhe-2A (Milky Way-2A), a system developed by China’s National University of Defense Technology (NUDT) and deployed at the National Supercomputer Center in Guangzho, China retained the No. 5 system with 61.4 Pflop/s.
HPC5 at #6 is the second new system. It is a PowerEdge system built by Dell installed by the Italian company Eni S.p.A. and is now the most powerful system in Europe. It achieves its performance of 35.5 Pflop/s by using NVIDIA Tesla V100 as accelerators and Mellanox HDR Infiniband as network.
Selene at #7 is an NVIDIA DGX A100 SuperPOD installed in house at NVIDIA in the USA. Naturally it uses NVIDIA Ampere A100 for acceleration and a Mellanox HDR Infiniband as network as well to achieve it’s 27.6 Pflop/s.
Frontera, a Dell C6420 system was installed at the Texas Advanced Computing Center of the University of Texas earlier this year and is listed at No. 8. It achieved 23.5 Pflop/s using 448,448 of its intel Xeon cores.
Marconi-100 at #9 is the second Italien system in the Top10. It is an IBM Power System AC922 based on IBM POWER9 processors, Nvidia Volta V100 accelerators, and a Dual-rail Mellanox EDR Infiniband. The system is installed at the Italien research center CINECA and achieved 21.6 Pflop/s.
The No. 10 is the Piz Daint, a Cray XC50 system installed at the Swiss National Supercomputing Centre (CSCS) in Lugano, Switzerland.
A total of 146 systems on the list are using accelerator/co-processor technology, up from 145 six months ago. 1 of these use NVIDIA Ampere chips, 107 use NVIDIA Volta, and 0 systems with 17.
Intel continues to provide the processors for the largest share (94.20 percent) of TOP500 systems.
We have incorporated the HPCG benchmark results into the Top500 list to provide a more balanced look at performance.
Supercomputer Fugaku takes the leadership followed by the 2 top DOE systems Sierra and Summit in the #2 and #3 spots with reespect to HPCG performance.
Japanese systems continue to take leading roles in the Green500 and take the first spot with MN-3, a Preferred Networks system with 21.1 GFlops/watts.
The entry level to the list moved up to the 1.23 Pflop/s mark on the Linpack benchmark.
The last system on the newest list was listed at position 450 in the previous TOP500.
Total combined performance of all 500 exceeded the Exaflop barrier with now 2.21 exaflop/s (Eflop/s) up from 1.65 exaflop/s (Eflop/s) 6 months ago.
The entry point for the TOP100 increased to 2.80 Pflop/s.
The average concurrency level in the TOP500 is 142,320 cores per system up from 126,308 six months ago.
The most energy-efficient system and No. 1 on the Green500 is a new MN-Core Server from Preferred Networks installed at Preferred Networks , Japan. It achieved 21.1 GFlops/Watt power-efficiency during its 1.62 Pflop/s Linpack performance run. It is listed on position 395 in the TOP500.
In second position is the new NVIDIA system Selene, a DGX A100 SuperPOD system powered by an AMD 7742 processor. It is on position 7 in the TOP500.
In third position is the NA-1 system, a PEZY Computing / Exascaler Inc. system installed at NA Simulation in Japan. It achieves 18.4 GFlops/Watt energy efficiency. It is on position 470 in the TOP500.
Supercomputer Fugaku is now the leader on the HPCG benchmark with 13.4 PFlop/s.
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.