For the third time in a row, Frontier is ranked primary among the many world’s quickest supercomputers, and it stays the one whose quickest velocity exceeds one exaFLOPS.
At 1.194 quintillion floating level operations per second (FLOPS) Frontier stored its rating with greater than double the highest velocity of its nearest competitor, in response to the record compiled by TOP500, which points the rankings twice a yr. A quintillion is 1018 or one exaFLOPS (EFLOPS).
The quantity two machine, Fugaku, maxed out at 442.01petaFLOPS. A petaFLOPS is 1015 FLOPS.
Two opponents within the prime 10 improved their speeds for the reason that final rating interval that resulted in November 2022, however not practically sufficient to even draw shut. These two—LUMI and Leonardo—rank third and fourth, respectively.
The ten quickest on the record are the identical ones as final time and rank in the identical order.
The velocity measurement utilized in evaluating the computer systems is called the Excessive Efficiency Linpack (HPL) benchmark, which measures how effectively methods remedy a dense system of linear equations.
Half of the highest 10 are deployed within the US. Two are in China, and one every in Finald, Italy, and Japan.
Along with rating primary for flat-out velocity, Frontier additionally ranked primary in suitability for doing synthetic intelligence (AI) capabilities, a classification referred to as HPL-MxP.
Frontier and LUMI had been additionally among the many prime 10 for energy effectivity as measured by GFLOPS per Watt. Frontier got here in at quantity six with 52.592GFLOPS/W, and LUMI got here in quantity seven with 51.382GFLOPS/W.
Probably the most energy environment friendly total was Henri, operated by the Flatiorn Institute of New York Metropolis, with 65.396 GFLOPS/W.
Listed here are the ten total quickest on the TOP500 record for June 2023 and the place they’re situated:
#1—Frontier
An HPE Cray EX system run by the US Division of Vitality, Frontier incorporates third Gen AMD EPYC™ CPUs representing 8,730,112 cores which were optimized for high-performance computing (HPC) and AI with AMD Intuition™ 250X accelerators and Slingshot-11 interconnects. Its HPL benchmark was 1.194EFLOPS.
#2—Fugaku
Supercomputer Fugaku, housed on the RIKEN Middle for Computational Science in Kobe, Japan, scored 442.01PFLOPS within the HPL take a look at. It’s constructed on the Fujitsu A64FX microprocessor and has 7,630,848 cores.
#3—LUMI
LUMI is an HPE Cray EX system on the EuroHPC heart at CSC in Kajaani, Finland, with a efficiency of 309.1 PFLOPS. It depends on AMD processors and boasts 2,220,288 cores.
#4—Leonardo
Leonardo, which resides in Bolbogna, Italy, is an Intel/Nvidia system with 1,463,616 cores and a most velocity of 238.7PFLOPS.
#5—Summit
An IBM system at Oak Ridge Nationwide Laboratory in Tennessee, Summit scored 148.6PFLOPS on the HPL benchmark. It has 4,356 nodes, every with two Power9 CPUs with 22 cores and 6 Nvidia Tesla V100 GPUs, every with 80 streaming multiprocessors (SM). The nodes are linked by a Mellanox dual-rail EDR InfiniBand community. It has 2,414,592 cores.
#6—Sierra
Related in structure to Summit, Sierra reached 94.64 PFLOPS. It has 4,320 nodes with two Power9 CPUs and 4 Nvidia Tesla V100 GPUs and a complete of 1,572,480 cores. It’s housed on the Lawrence Livermore Nationwide Laboratory, in California.
#7—Sunway TaihuLight
Sunway TaihuLight is a machine developed by Nationwide Analysis Middle of Parallel Laptop Engineering & Expertise (NRCPC) in China and is put in within the metropolis of Wuxi. It reached 93.01PFLOPS on the HPL benchmark. It has 10,649,600 cores.
#8—Perlmutter
The Perlmutter system is predicated on the HPE Cray Shasta platform and is a heterogeneous system with each AMD EPYC-based nodes and 1536 Nvidia A100-accelerated nodes. It has 761,856 cores. It achieved 70.87 PFLOPS. That’s an enchancment of about 6PFLOP/s over final yr’s rating, however nonetheless not sufficient to catch Sunway TaihuLight.
#9—Selene
Selene is an Nvidia DGX A100 SuperPOD based mostly on an AMD EPYC processor with Nvidia A100s for acceleration and a Mellanox HDR InfiniBand as a community. It has 555,520 cores. It achieved 63.46 PFLOPS and is put in in-house at Nvidia services within the US.
#10—Tianhe-2A (Milky Approach-2A)
Powered by Intel Xeon CPUs and NUDT’s Matrix-2000 DSP accelerators, Tianhe-2A has 4,981,760 cores within the system to attain 61.44 PFLOPS. It was developed by China’s Nationwide College of Protection Expertise (NUDT) and is deployed on the Nationwide Supercomputer Middle in Guangzhou, China.
Copyright © 2023 IDG Communications, Inc.
