Network Security Internet Technology Development Database Servers Mobile Phone Android Software Apple Software Computer Software News IT Information

In addition to Weibo, there is also WeChat

Please pay attention

WeChat public account

Shulou

The world's top 500 supercomputers have announced that Frontier is still the only 10 billion-level supercomputer on earth.

2025-03-29 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > IT Information >

Share

Shulou(Shulou.com)11/24 Report--

CTOnews.com Oak Ridge National Laboratory's (ORNL) Frontier remains the only exascale machine on Earth, and its HPL score has increased from 1.02 Eflop/s in November 2022 to 1.194 Eflop/s today, an increase of about 17%, according to the 61-issue Supercomputer TOP500 released today.

Additionally, Frontier achieved a score of 9.95 Eflop/s on the HLP-MxP benchmark, which measures hybrid precision computational performance. This is also an improvement over the 7.94 Eflop/s achieved by the system in the previous list and nearly 10 times higher than the machine's HPL score.

Frontier is based on the HPE Cray EX 235a architecture and powered by AMD EPYC processors (64C 2GHz) with 8699,904 cores and an incredible 52.59 Gflops / watt efficiency rating (sixth on the GREEN500).

In addition to Frontier, Fugaku Fuyue, located at the Riken Computational Science Center (R-CCS) in Kobe, Japan, remained second in the list, unchanged from the previous HPL score of 0.442 Eflop/s.

Finland's EuroHPC / CSC LUMI system entered the list at 3rd place in June 2022 and remained at 3rd place after an upgrade last November with an HPL score of 0.3091 Eflop/s. As a result, it remains the largest supercomputer system in Europe.

EuroHPC / CINECA's Leonardo system in Bologna, Italy, remains in fourth place. It was also upgraded, from 0.174 Eflop/s to 0.239 Eflop/s.

Fifth place remains IBM's Summit system built at Oak Ridge National Laboratory (ORNL) in Tennessee, which maintains its fifth place performance on the HPL benchmark at 148.8 Pflop/s.

Sierra, a system at Lawrence Livermore National Laboratory in California, is ranked No. 6. Its architecture is very similar to Summit, the fifth-ranked system, with 4,320 nodes, two POWER9 CPUs and four NVIDIA Tesla V100 GPUs, and a performance of 94.6 Pflop/s.

Sunway TaihuLight, developed by China's National Research Center for Parallel Computer Engineering (NRCPC) and installed at China's National Supercomputing Center in Wuxi City, Jiangsu Province, ranked seventh with 93 Pflop/s.

Perlmutter, ranked 8th, is based on the HPE Cray"Shasta" platform and a heterogeneous system with AMD EPYC-based nodes and 1,536 NVIDIA A100 accelerated nodes with performance of 64.6 Pflop/s

Selene is the NVIDIA DGX A100 SuperPOD built in-house by NVIDIA. The system is based on AMD EPYC processors, equipped with NVIDIA A100 accelerators and Mellanox HDR InfiniBand as a network, and has a performance of 63.4 Pflop/s.

Tianhe-2A (Yinhe-2A) is a system developed by the National University of Defense Technology (NUDT) and deployed at the National Supercomputer Center in Guangzhou, China, currently ranked 10th at 61.4 Pflop/s.

According to the TOP500 list, AMD, Intel and IBM processors are the top choices for HPC. Four of the TOP10 supercomputing systems use AMD processors (Frontier, LUMI, Perlmutter, and Selene), two use Intel processors (Leonardo and Tianhe-2A), and two use IBM processors (Summit and Sierra).

As with previous lists, China and the United States dominate the entire TOP500 list. The U.S. lead increased from 126 units in the previous period to 150 units in this period, while China dropped from 162 units to 134 units. Asia as a whole has 192 systems, North America has 160 systems, and Europe has 133 systems.

Ethernet remained the clear winner in terms of system interconnect (although it dropped from 233 machines to 227);Infiniband interconnect increased from 194 to 200;Omnipath dropped from 36 to 35; and custom interconnect technology increased significantly from 4 to 31.

HenriCTOnews.com notes that the Henri system at Lenovo Flatiron Institute in New York, USA, once again topped the GREEN500 list with an astonishing energy efficiency of 65.40 Gflops / Watt. More importantly, it has risen from 405th to 255th on the TOP500 list after upgrading, and its current HPL score is 2.88 Pflop/s, up from 2.038 Pflop/s last time.

In second place is ORNL's Frontier Test and Development System (TDS), which has an efficiency rating of 62.20 Gflops / Watt and an HPL score of 19.2 Pflop/s.

In third place is the Adastra system, also known as the HPE Cray EX 235a system, based on AMD EPYC and AMD Instinct MI250X.

Frontier also deserves credit for energy efficiency. Not only does it rank first on the TOP500 list with an HPL score of 1.194 Gflops/ s, it also ranks sixth on the GREEN500 with an energy efficiency rating of 52.59 Gflops / Watt. The HPL performance of these systems demonstrates that high performance does not have to come at the expense of high power consumption.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.

Views: 0

*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.

Share To

IT Information

Wechat

© 2024 shulou.com SLNews company. All rights reserved.

12
Report