Here sharing with you the list of World’s Fastest Supercomputers 2021,you can easily understand the features and usease of these:
Fugaku: – This Japanese computer is considered to be the highest supercomputer that has ever been built with an Arm A64FX capability 7,630,848 cores. With the help of additional technology, Fugaku grew its HPL performance to 442 petaflops following its launch in June of this year. (This is three times more than the second-highest performing system on the rankings). Fugaku was built in the company of Fujitsu and is currently in use within the RIKEN Center for Computational Science (R-CCS) in Kobe, Japan. It is currently being used on an experimental basis in COVID-19 research.
Fugaku is the first top-rated system run by ARM processors. Two other new features are hybrid memory cubes that are connected to the processors, and a brand new version of the Tofu network, which provides seamless integration among all the nodes of the system. The designers see its success as evidence that there is room for further innovation in HPC.
Summit: This IBM-built machine located at the Oak Ridge National Laboratory (ORNL) located in Tennessee is the most powerful technology within the US with a speed in the range of 148.8 petaflops. Summit is home to 4,356 nodes each with the 22 core Power9 CPUs as well as Six NVIDIA Tesla V100 GPUs. It is part of the Department of Energy’s commitment to supercomputing power the Oak Ridge National Laboratory’s (ORNL) Summit first claimed the number one spot in June 2018 receiving the top spot over China to claim the top spot for the first time within six years. For a highly-performing computer, Summit also initially ranked third on the GreenTop500 that measures the efficiency of supercomputers in terms of energy consumption, however; it has since fallen to ninth.
Sierra: – Sierra located at the Lawrence Livermore National Laboratory in California is equipped with an HPL rating that is 94.6 petaflops. The architecture of Sierra is similar to the one of Summit, with each of the 4,320 nodes having two Power9 CPUs as well as the four NVIDIA Tesla V100 GPUs. The Lawrence Livermore National Lab’s (LLNL) Sierra initially debuted as #3 in the June 2018 list having 71.6 petaflops. The process was optimized later, bringing the speed of its 1,572,480 cores, to 94.6 petaflops, which earned it the second position in November of 2018. But the rise to the number one spot in June 2020 will push Sierra back to third place. With the two IBM (CPUs) along with NVIDIA graphic processing units (GPUs), It is specifically designed to model and simulations that are most required of The US National Nuclear Security Administration.
Sunway TaihuLight: – Technology was developed by the Chinese National Research Center of Parallel Computer Engineering & Technology (NRCPC) and is housed in the National Supercomputing Center. It is powered solely with Sunway SW26010 processors and can achieve an HPL performance of 93 petaflops. It was also a previous top-ranked machine; Sunway TaihuLight dominated the list for two years following its launch in June 2016. Its 93.01 petaflops and 10649,000 cores created the world’s most powerful supercomputer by a large distance, with more than 5 times as much processing capacity as its closest rival (ORNL’s Titan) and nearly 19 times the number of cores. The supercomputer is housed within the National Supercomputing Center in Wuxi, China; TaihuLight’s creators have utilized the supercomputer for years to tackle tasks that span the field of science to climate change to the science of climate and advanced production. It also has found the ability to forecast marine conditions aiding ships to avoid rough seas, as well as assisting the offshore drilling of oil.
Selene: – Selene NVIDIA DGX A100 SuperPOD has been house-built at NVIDIA Corporation. It was ranked at number seven in June. However, it has doubled in size and is now able to climb to the fifth spot. The system is built of AMD EPYC processors with NVIDIA’s A100 GPUs, which are the latest GPUs designed for acceleration. Selene has achieved 63.4 petaflops of HPL due to the upgrade.
Tianhe-2A: – It is also called the Milky Way-2A; this system is designed in the Chinese National University of Defense Technology (NUDT) and is installed within NUDT’s National Supercomputer Center in Guangzhou. It can reach 61.4 petaflops HPL. Tianhe-2 which means “MilkyWay-2,” originally debuted as the top-ranked computer in the world in June 2013. However, despite improvements throughout the years, 4,981,760 cores run with 61.4 petaflops.
It’s barely securing an appearance among the top five. This is the ephemeral beauty of a modern supercomputer. TOP500 stated that the machine was created in collaboration with NUDT, the National University of Defense Technology (NUDT) in China, and is designed primarily for security-related applications in the government sector. This means that the majority of the work carried out by Tianhe-2 remains secret however If its processing capabilities are an indication it’s likely to be involved in some significant projects.
JUWELSBooster Module: – The JUWELS Booster Module is a debutant in the lineup, the built by Atos BullSequana machine was installed recently at the Researchszentrum Julich (FZJ) situated in Germany. It’s part of a modular structure. These modules are linked making use of ParTec Modulo Cluster Software Suite. ParTec modulo Cluster Software Suite. This Booster Module uses AMD EPYC processors and NVIDIA A100 GPUs that provide acceleration comparable to the sixth-ranked Selene system. By itself, this JUWELS Booster Module has been able to reach 44.1 HPL petaflops. This is the highest-powered device in Europe.
HPC5: – A Dell Power Edge system, it’s installed in the Italian company Eni S.p.A. Easily can achieve a speed of 35.5 petaflops, using Intel Xeon Gold CPUs and NVIDIA Tesla V100 GPUs. The Tesla V100 is among the fastest machine on the list that is employed for commercial uses at a customer’s website.
Frontera: – A Dell C6420 system was put in place in the Texas Advanced Computing Center of the University of Texas in 2019. It can achieve 23.5 petaflops by using 448,448 of its Intel Platinum Xeon cores.
Dammam-7: – The new supercomputer is located in Saudi Aramco in Saudi Arabia and is the second supercomputer commercially owned that is currently in the top 10 list. Its HPE Cray CS-Storm systems use Intel Gold Xeon CPUs and NVIDIA Tesla V100 GPUs. It was able to reach 22.4 petaflops when tested on HPL.
