The US Department of Energy's National Nuclear Security Administration (DOE/NNSA) has signed a USD600 million contract with Cray Inc to build its first exascale supercomputer, El Capitan. The computer will be housed at the Lawrence Livermore National Laboratory (LLNL) in California.

(Image: DOE)

Exascale supercomputers are capable of executing a quintillion (1018) calculations each second. This is much faster than the fastest supercomputers in the world today which solve problems at the petascale, or a quadrillion (1015) calculations per second. According to the Exascale Computing Project -  a collaborative effort of the DOE's Office of Science and the NNSA - exascale computers represent the next milestone in computing achievement. Their ability to simulate more realistically the processes involved in a range of areas including precision medicine, climate modelling, materials science and nuclear physics will have profound impacts on everyday life, it says.

"The Department of Energy is the world leader in supercomputing and El Capitan is a critical addition to our next generation systems," US Energy Secretary Rick Perry said. "El Capitan's advanced capabilities for modelling, simulation, and Artificial Intelligence will help push America's competitive edge in energy and national security, allow us to ask tougher questions, solve greater challenges, and develop better solutions for generations to come." 

El Capitan will be used to perform "mission critical" research to maintain the US nuclear weapons stockpile, NNSA said. "It will be used by researchers … to run 3D simulations and calculations at resolutions that are difficult, time-consuming, or even impossible using today's state-of-the art supercomputers," the administration added.

The computer, which is scheduled for delivery in late 2022, will be the DOE's third exascale supercomputer after Aurora at Argonne National Laboratory and Frontier at Oak Ridge National Laboratory (ORNL), both of which are scheduled for delivery in 2021. All three are to be built by Cray using their Shasta architecture, Slingshot interconnect, and new system software platform.

ORNL's Summit supercomputer in 2018 was recognised as the fastest supercomputer in the world, when it reached a speed of 122.3 petaflops, which it has since improved to 148.6 petaflops, according to the semi-annual TOP500 List ranking of the world's fastest computing systems. Second place is currently held by the Sierra supercomputer at LLNL, with a fastest speed of 94.6 petaflops. One petaflop is equal to one quadrillion calculations - or "floating-point operations" - per second.

Researched and written by World Nuclear News

Date: Thursday, 15 August 2019
Original article: