Top 10 Most Powerful Supercomputers | Jumpstart

by Assessor
0 comment

Japan’s Fugaku supercomputer has retained its place because the primary laptop system, adopted by IBM’s Summit and Sierra within the U.S.

Within the 56th version of the TOP500, Japan, USA, Germany, Italy and France have dominated the highest 100 of the world’s strongest supercomputers. Japan’s Fugaku supercomputer retained its place because the primary laptop system, adopted by IBM’s Summit and Sierra within the U.S.

See more: Top 10 fastest supercomputer in the world

The Top500 venture tracks essentially the most highly effective supercomputers on the planet and is printed biannually. The non-profit group launched its November 2020 outcomes final week. This 12 months, whereas two new programs made it to the highest 10, the complete listing recorded the smallest variety of new entries for the reason that venture’s inception in 1993.

Because of this, Top500 representatives urged that the present listing displays a flattening efficiency development curve.

What the outcomes present

The world’s supercomputers at the moment are not solely sooner, but in addition smarter, and assist a higher number of workloads.

Practically 70% (140) of the programs on the TOP500 listing, together with eight of the highest 10, are powered by NVIDIA know-how and are “more and more utilizing AI to assist researchers make discoveries sooner.”

NVIDIA’s end-to-end HGX AI supercomputing platform helps speed up scientific computing, information analytics, and AI workloads.

“The shift to incorporating AI into HPC (Excessive-Efficiency Computing), and a platform that extends past conventional supercomputing facilities, represents a big change in a subject that, since Seymour Cray’s CDC 6600 was launched in 1964, has centered on harnessing ever bigger, extra highly effective machines for compute-intensive simulation and modeling,” NVIDIA stated in a weblog submit.

In October this 12 months, Italian inter-university consortium CINECA had introduced its plans to construct the world’s quickest AI supercomputer. The brand new “Leonardo” system is anticipated to ship 10 exaflops of FP16 efficiency in AI workloads. It is going to function round 14,000 Nvidia A100 GPUs and round 3,500 Intel Sapphire Rapids CPUs.

Within the case of a supercomputer, its efficiency capabilities are indicated utilizing a typical charge known as FLOPS or floating-point operations per second. This charge signifies the variety of floating-point arithmetic calculations a system can carry out per second. Whereas a 1 petaFLOPS (PFLOPS) laptop system can carry out one quadrillion (10^15) floating-point operations per second, 1 exaFLOPS (EFLOPS) laptop system can carry out one quintillion (10^18) floating-point operations per second.

In the meantime, Intel has remained the popular processor supplier, with 90% of the programs outfitted with its Xeon or Xeon Phi chips.

With 212 programs on the listing, China has exerted unequivocal domination over the listing, adopted by the U.S. with 113 programs, and Japan with 34. Nevertheless, on the subject of combination efficiency, the U.S. continues to guide at 668.7 petaflops, in comparison with China’s 564.0 petaflops. In the meantime, Japan has an combination efficiency of 593.7 petaflops.

Supercomputers, which have the unlucky aspect impact of consuming monumental quantities of power, at the moment are additionally ranked on the Green500 by power effectivity. On the Green500 listing, the brand new NVIDIA DGX SuperPOD within the U.S. was discovered to be essentially the most energy-efficient system. It was adopted by MN-3 in Japan, and Atos-built JUWELS Booster Module in Germany.

Prime 10 supercomputers

Notable developments within the prime 10 embody the addition of two new programs, and a brand new efficiency benchmark set by the Fugaku supercomputer.

As a result of addition of latest {hardware} to beef up the system, Fugaku grew its Excessive Efficiency LINPACK (HPL) benchmark to 442 petaflops, in comparison with the 416 petaflops the system achieved when it debuted in June 2020 – a 6.4% velocity enchancment.

Editor’s Choice: 15 Most Evil People In History | 15 Most Notorious Men In History

The HPL benchmark, which signifies the efficiency of a devoted system when fixing a dense system of equations, is used to find out the Top500.

1. Fugaku, Japan

Constructed by Fujitsu, Fugaku is put in on the RIKEN Heart for Computational Science (R-CCS) in Kobe, Japan. With its further {hardware}, the system achieved a brand new world file of 442 petaflops consequence on HPL, making it thrice forward of the quantity two system within the listing.

RIKEN’s director, Satoshi Matsuoka, acknowledged that the advance got here as they had been “lastly with the ability to use your entire machine fairly than only a good chunk of it.”

For the reason that June competitors, his group has been capable of fine-tune the code for max efficiency. “I don’t assume we are able to enhance a lot anymore,” Matsuoka stated.

2. Summit, U.S.

Based mostly on the Oak Ridge Nationwide Laboratory (ORNL) in Tennessee, Summit was constructed by IBM and is the quickest system within the US.. Launched in 2018, it has a efficiency of 148.8 petaflops and has 4,356 nodes, every one housing two 22-core Power9 CPUs and 6 NVIDIA Tesla V100 GPUs.

Not too long ago, two groups engaged on Summit received the celebrated Gordon Bell Prize for excellent achievement in high-performance computing, generally known as the ‘Nobel Prize of supercomputing.’

3. Sierra, U.S.

A system on the Lawrence Livermore Nationwide Laboratory (LLNL) in California, Sierra has an HPL mark of 94.6 petaflops. With every of its 4,320 nodes outfitted with two Power9 CPUs and 4 NVIDIA Tesla V100 GPUs, it has an structure much like that of Summit.

Sierra additionally made it to the fifteenth place on the Green500 Checklist of the world’s most energy-efficient supercomputers.

4. Sunway TaihuLight, China

Put in at China’s Nationwide Supercomputing Heart in Wuxi, Sunway TaihuLight beforehand held the Number one spot for 2 years (2016-2017). Nevertheless, its rank has since fallen. Whereas it was in third place final 12 months, it has now slipped to fourth.

Constructed by China’s Nationwide Analysis Heart of Parallel Pc Engineering & Expertise (NRCPC), it achieved 93 petaflops on its HPL benchmark. It’s powered solely by Sunway SW26010 processors.

5. Selene, U.S.

Put in in-house at NVIDIA Corp, Selene jumped to fifth place from seventh place within the June rankings. After a current improve, Selene achieved 63.4 petaflops on HPL, almost doubling its earlier rating of 27.6 petaflops.

Editor’s Choice: Top 15 Ice Cream Brands in the world – Worlds top Ice Creams

NVIDIA unveiled Selene, its AI supercomputer, in June this 12 months, after setting up and working it in lower than a month. Its key makes use of embody system improvement and testing, in-house AI workloads, and chip design work.

6. Tianhe-2A, China

With 61.4 petaflops, Tianhe-2A (aka MilkyWay-2A) climbed one spot to sixth place. Developed by China’s Nationwide College of Protection Expertise (NUDT), it’s put in on the Nationwide Supercomputer Heart in Guangzhou.

Tianhe-2A is powered by Intel Xeon CPUs and NUDT’s Matrix-2000 DSP accelerators. It is going to be used for simulation, evaluation, and authorities safety purposes. It held the Number one spot from June 2013 to November 2015.

7. JUWELS Booster Module, Germany

Atos-built JUWELS Booster Module is the latest entrant on the listing. The BullSequana machine was lately put in on the Forschungszentrum Jülich (FZJ) in Germany and is essentially the most highly effective system in Europe, coming in at 44.1 HPL petaflops.

Based mostly on a modular system structure, JUWELS is powered by AMD processors and NVIDIA GPUs, much like the Selene system.

8. HPC5, Italy

A Dell PowerEdge system, HPC5 was put in by the Italian firm Eni S.p.A and is situated inside Eni’s Inexperienced Knowledge Heart in Italy. Some of the highly effective and sustainable computing programs on the planet, HPC5 is used to analysis new sources of power.

With a efficiency of 35.5 petaflops, it’s the strongest system within the listing used for industrial functions at a buyer website. It makes use of Intel Xeon Gold CPUs and NVIDIA Tesla V100 GPUs.

9. Frontera, US

Put in on the Superior Computing Heart on the College of Texas at Austin in September 12 months, Frontera is a Dell C6420 system outfitted by Intel. Utilizing 448,448 of its Intel Platinum Xeon cores, the system achieves 23.5 petaflops.

Frontera aids analysis throughout all areas of science together with these associated to quantum mechanics, drug design, eradicating rising viruses, and the physics of black holes.

10. Dammam-7, Saudi Arabia

Dammam-7 is the second new addition on the prime of the listing. Put in at Saudi Aramco in Saudi Arabia, the HPE Cray CS-Storm programs makes use of Intel Gold Xeon CPUs and NVIDIA Tesla V100 GPUs. It achieves 22.4 petaflops and is the second industrial supercomputer within the prime 10.

Picture by Brett Sayles from Pexels

Editor’s Choice: What are the Largest Film Festivals in the World? – Mack Sennett Studios

Related Posts

Leave a Comment