Powered by AMD EPYC processors and AMD Instinct accelerators, the HPE Cray EX supercomputer delivered 1.1 exaflops of performance to beat the Japan-based Supercomputer Fugaku to secure the no.1 ranking in the latest Top500 and Green500 lists
The Frontier supercomputer at the Oak Ridge National Laboratory. (Credit: Oak Ridge National Laboratory)
Frontier, a new supercomputer built by Hewlett Packard Enterprise (HPE) for the US Department of Energy (DOE)’s Oak Ridge National Laboratory (ORNL), has broken the exascale speed barrier to become the fastest supercomputer in the world.
Powered by AMD EPYC processors and AMD Instinct accelerators, the HPE Cray EX supercomputer delivered 1.1 exaflops of performance to take the no.1 ranking in the latest Top500 and Green500 lists.
The performance of the Frontier supercomputer is claimed to be more than two times the number two system, which is the Supercomputer Fugaku installed at the RIKEN Center for Computational Science in Japan.
Furthermore, the ORNL supercomputer’s performance is more than the combined performance of the next seven systems on the latest Top500 list.
It also ranked number one in the mixed-precision computing category which rates performance in formats that are used commonly for artificial intelligence. The supercomputer took the no.1 ranking in the category with a performance of 6.88 exaflops.
HPE said that the Frontier supercomputer will enable scientists in modelling and simulating at an exascale level. This will help in solving problems that are eight times more complex, up to ten times faster than today’s leading supercomputers.
Installed in late 2021, Frontier is also likely to achieve even greater levels of speed with a theoretical peak performance of two exaflops, said HPE.
ORNL computing and computational sciences associate lab director Jeff Nichols said: “When researchers gain access to the fully operational Frontier system later this year, it will mark the culmination of work that began over three years ago involving hundreds of talented people across the Department of Energy and our industry partners at HPE and AMD.
“Scientists and engineers from around the world will put these extraordinary computing speeds to work to solve some of the most challenging questions of our era, and many will begin their exploration on Day One.”
Frontier features 74 HPE Cray EX supercomputer cabinets, which are specifically built for supporting next-generation supercomputing performance and scale, once it is open for early science access.
Each of its nodes contains an optimised EPYC processor and four AMD Instinct accelerators to total over 9,400 central processing units (CPUs) and more than 37,000 graphics processing units (GPUs) in the entire machine.
The supercomputer involves 141km worth of HPE Slingshot networking cables.
It features the Cray Clusterstor E1000 storage system that enables the machine’s Orion storage system in delivering 700 petabytes of storage capacity. Orion also has peak write speeds of over 35 terabytes per second along with 15 billion random-read input/output operations per second.