Little-known Japanese CPU threatens to make Nvidia, Intel and AMD obsolete in HPC market

Sandia National Laboratories has announced it will be the first Department of Energy labs in the US to deploy the Fujitsu A64FX, the only ARM-based processor designed from ground up for HPC projects and supercomputers.

Fujitsu is known primarily for its business laptops, tablets and desktops, but is a behemoth in its own right when it comes to processors, having been in the business for well over half a century.

Launched in 2019, the CPU has 48 cores, a theoretical peak performance of 3.38 TFLOPS, runs at 2.2GHz and has 32GB HBM2 memory on the die itself.

Check out our list of the best small business servers of 2020
Here's our list of the best dedicated server hosting services around
We've built a list of the best Windows server hosting services available

What makes it ideal for the HPC market is that it provides far higher bandwidth performance between memory and the CPU - up to 1TBps. Moving data to and from the CPU is the biggest obstacle by far to what researchers refer to as exascale computing.

What makes the A64FX even more exciting is that Fujitsu wants the technology to trickle down to hyperscalers and major cloud computing giants so that the masses can benefit too.

Given it is based on ARM architecture, it can (and has) run Linux distributions out of the box and even Microsoft Windows.

It is considered a general purpose CPU, but surpasses even GPUs from Nvidia and AMD on the all-important metric of performance per watt. Indeed, a 768-CPU prototype sits on top of the Green500 list - the leaderboard for supercomputers that deliver the most power per watt.

The A64FX was designed expressly to power the successor of Japan’s main supercomputer, the K, which was decommissioned back in August 2019.

Its replacement - the Fugaku - is expected to be 100 times faster when it launches later this year, will run on a Linux distribution called McKernel and will reach a staggering 400 petaflops. The aim is for it to be the first supercomputer to hit one exaflop when fully deployed with half a million processors buzzing.