Frontier, the HPE Cray exa-scale supercomputer operated by the Department of Energy (DoE) at Oak Ridge National Laboratory (ORNL), remains by far the fastest supercomputer on the planet, according to the Top 500.
With a High-Performance Linpack (HPL) benchmark score of 1.102 exaflops per second, Frontier is the only exa-scale supercomputer and is three times faster than its closest competitor, Fugaku, at the RIKEN Center for Computational Science ( R-CCS) in Kobe, Japan, according to Top 500. An exaflop is a quintillion calculations or floating point operations (flops) per second.
Frontier, an HPE Cray EX system, is equipped with 8,730,112 cores and combines 3rd generation AMD EPYC processors optimized for HPC and AI, with AMD Instinct 250X accelerators and a Slingshot-10 interconnect.
Frontier secured the top spot in the Top 500 for June, when it achieved the same performance of 1.1 exaflops. It has a theoretical peak of 2 exaflops and is 10 times more potent than ORNL’s Summit system.
Also: White House Passes AI ‘Bill of Rights’ That Tries to Allay Your Concerns
Frontier is also part of the DoE’s Exacscale Computing project. According to ORNL, it’s housed in 74 behemoth HPE Cray EX supercomputer cabinets, which include more than 9,400 AMD-powered nodes and 90 miles of network cables. It also has 37,000 GPUs.
As ORNL explains, modeling cancer cells, supernovas, the coronavirus, or the atomic structure of elements requires the 64-bit precision offered by typical supercomputers, whereas machine learning algorithms require less precession. , for example 32, 24 or 16 bit precision and can therefore take advantage of GPUs.
Frontier’s Orion storage system holds 700 petabytes of data; each cabinet weighs 8,000 pounds and the system is cooled by 6,000 gallons of water pumped every minute by four 350 horsepower pumps. The pumps could fill an Olympic-size swimming pool in 30 minutes. “Frontier is the clear winner of the exascale race, and it will take a lot of hard work and innovation to bring it down from the top spot,” said the Top 500 press release.
The fastest supercomputer in Europe, in third place in the Top 500, is the upgraded LUMI system based in Finland, also an HPE Cray EX system. It is installed at the EuroHPC center at CSC’s data center in Kajaani, Finland, and has a performance of 309.1 Pflop/s.
As Top 500 notes, the European High Performance Computing Joint Undertaking (EuroHPC JU) is pooling European resources to develop high-end exascale supercomputers for big data processing.
A newcomer is Leonardo, installed on the EuroHPC site of Cineca, in Italy. It is a French-made Atos BullSequana XH2000 system with Xeon Platinum 8358 32C 2.6 GHz as main processors, NVIDIA A100 SXM4 40 GB as accelerators and Quad-rail NVIDIA HDR100 Infiniband as interconnect. It achieved a Linpack performance of 174.7 Pflop/s.
Atos unveiled the XH300 earlier this year, a more powerful system than the XH200 with processors, GPUs and accelerators from AMD, Intel and Nvidia, as well as microprocessor technology from French company The Silicon Pearl. (SiPearl).
Old Summit, a system built by IBM, ranks fifth with a performance of 148.8 Pflop/s.
Sierra, Summit’s sister at Lawrence Livermore National Laboratory in California, is in sixth place with a score of 94.6 Pflop/s.
China’s fastest supercomputer, Sunway TaihuLight, was the world’s fastest supercomputer in 2017, but now ranks seventh with a performance of 93 Pflops/s. Tianhe-2A (Milky Way-2A) was the second fastest in 2017 but now ranks 10th with a performance of 61.4 Pflop/s. They were eclipsed by Summit and Sierra in 2018.
While China’s fastest supercomputers have slipped in the HPL Top 500 rankings, China still dominates the Top 500 list along with the United States. The United States has 126 machines in the Top 500, while China has 162 in this ranking. However, the number of Chinese systems fell from 173 in the June 2022 ranking. Europe as a whole had 131 machines in the Top 500, down from 118 machines in June.
#giant #supercomputer #dominates #leaderboard #catching #lot #work