A supercomputer jointly developed by Riken and Fujitsu using ARM technology is now ranked as the world’s fastest.
The latest biannual Top 500 supercomputer ranking has put a machine called Fugaku in the top spot as the fastest of its kind in the world. Based in the city of Kobe and co-developed by Riken and Fujitsu, the supercomputer is the first built on ARM processors to reach the top spot in the ranking and is the first Japanese machine to do so since 2011.
Having achieved 415.53 petaflops, Fugaku is almost three times as fast as IBM’s Summit supercomputer, which clocks in at 148.6 petaflops. The Japanese supercomputer also topped the list in terms of high-performance conjugate gradient, where it scored 13,400 teraflops using 138,240 nodes, and in terms of the HPL-AI benchmark.
In testing using 92,160 nodes, Fugaku solved a breadth-first search of a graph with 1.1trn nodes and 17.6trn edges in approximately 0.25 seconds. This earned it a score of 70,980 gigaTEPS, more than doubling the score of 31,303 gigaTEPS achieved by Riken’s K computer and surpassing China’s Sunway TaihuLight, which is currently second on the Graph 500 list with 23,756 gigaTEPS.
#Fugaku has become No.1 in all the supercomputer performance benchmarks, #Top500, #HPCG, HPL-AI, and the #Graph500 for the first time in history as a single machine simultaneously. Thanks for putting up the list! https://t.co/iM1o0gjrtZ
— Satoshi Matsuoka (@ProfMatsuoka) June 22, 2020
Fujitsu’s corporate executive officer, Naoki Shinjo, said the company was “particularly proud” that it was able to achieve the new record just one month after the delivery of the system was finished and during the Covid-19 crisis.
Rene Haas, president of ARM’s IP Products Group, said the supercomputer “illustrates a dramatic shift in the type of compute that has been traditionally used in these powerful machines”.
The supercomputer is currently being used on an experimental basis for Covid-19 research, including on diagnostics, therapeutics and simulations of the spread of the virus. It’s expected to begin full operations from April of next year.
By this time, it’s hoped that it can run applications in areas such as drug discovery, energy creation, climate forecasting and “elucidation of the fundamental laws and evolution of the universe”.