It is the world’s largest supercomputer and its speed has broken all barriers.

Updates

It is the first supercomputer in history to have overcome the exascale barrier: it performs more than a trillion operations per second in a continuous manner.

Frontier, made with
Frontier is built on the latest generation AMD EPYC processors.EM

The Frontier System at Oak Ridge National Laboratory, a research center located in Tennessee, USA, has recently been crowned the world’s largest supercomputer. This is a great achievement in itself, but in this case we are also talking about the first supercomputer in history that has exceeded the capability of performing more than one. trillion operations per second in a continuous manner.

In computing, a flop is an abbreviation used to refer to floating point operations and the speed at which a computer can perform them, based on knowing its raw power and comparing large supercomputing systems with each other. allows for. One exaflop is equal to 10 to the power of 18 trades, i.e. over 1,000,000,000,000,000,000 of these calculations per second.

In synthetic performance tests, Frontier has managed to achieve 1.102 ExaFlops per second of sustained power, with peaks 1.69 ExaFlops per second, This computing power puts it far ahead of the world’s largest supercomputer: the Fugaku at the RIKEN Computer Science Center in Kobe, Japan.

The Fugaku has a computing power of 442 petaflops per second, but is also considered an exscale machine because its theoretical power allows it to reach one trillion operations per second, although this has not been achieved in practice.

Built on the latest generation of AMD EPYC processors and the AMD Instinct MI250x Compute Accelerator, the Frontier could theoretically exceed 2 ExaFlops per second of compute power for future changes.

Not only is he significantly more powerful than Fugaku, but he Has more computing power than the world’s seven largest supercomputers combined, It has also become the “greenest” supercomputer, topping the list of power per watt. It achieves a power efficiency of 62.68 gigaflops per watt consumed.

“Innovation and getting more performance and efficiency out of supercomputers is key to solving the world’s most complex challenges,” said Forrest Norrod, senior vice president and general manager of data center solutions at AMD.

One of the applications of this new system will be the study of the brain. When performing calculations mixed with the artificial intelligence-focused HPL-AI synthetic test battery (different and less intensive than the ones used to classify these machines), the Frontier is capable of reaching 6.88 exaflops, which is 68 million per second It is enough to emulate more than one instruction. Over 86,000 million neurons.

The computer is made up of 74 modules distributed over a room of 350 square meters. they live in Total 602,112 Process Cores and 4.6 petabytes of RAM. AMD Instinct adds MI250x acceleration units to the set 8,138,240 Additional Process Cores And the entire assembly is kept at an optimum temperature due to the continuously circulating liquid cooling system with 22,000 liters of coolant. During its typical operation the supercomputer consumes approx. 29 MW powerThat is, the estimated amount of energy required for 20,000 homes.

His rule, in any chaos, would be short. AMD is already working on another supercomputer, El Capitan, which will be operational in 2023 and will easily exceed 2 ExaFlops per second. The machine will form part of the system at Lawrence Livermore National Laboratory in California, which is also dependent on the US Department of Energy, such as Oak Ridge.

according to the norms of

trust project

know more


About the Author

The co-owner & marketing chief of "The Business News", Sravya is also good at Writing and communicating. she has good networking skills. she is really passionate about publishing quality news articles. - Thebusinessnews.org - You can reach Her at Facebook:- @sai.sravya.3910

Leave A Response