Japan’s Fugaku Is Again the World’s Fastest Supercomputer

Japans Fugaku Is Again the Worlds Fastest Supercomputer
Image Source:

A supercomputer has several thousand processor cores. It requires millions of dollars to build as well as maintain. Further, it needs an entire building to house the structure and cool it. Builders of supercomputers face a range of challenges. Despite this, several countries, like Japan, China, and the United States are racing to develop new and faster supercomputers. 

What is a supercomputer?

A supercomputer has a high level of performance, which is usually measured in floating-point operations per second (FLOPS). In 2017 some supercomputers can perform more than 1017 FLOPS. In addition, all of the 500 fastest supercomputers in the world run on Linux-based operating systems. Supercomputers were introduced in 1964, with the release of the CDC 6600.

Where do you use supercomputers?

Supercomputers are important in this day and age, as the need for computationally intensive tasks in different fields becomes more critical. Supercomputers perform various tasks, including molecular modeling, oil and gas exploration, quantum mechanics, cryptanalysis, climate research, weather forecasting. They are also used for physical simulations, such as the detonation of nuclear weapons, nuclear fusion, and spacecraft and airplane aerodynamics.

Currently, most of the supercomputers are united in various research for anti-Covid 19 vaccines and treatments. 

Fastest supercomputers

With the need for quick analysis of various data for various fields increases, supercomputers are expected to operate faster. There is ongoing competition among supercomputer builders on which one is the fastest.  

Currently holding first place again is Japan’s Fugaku, developed by Fujitsu-Riken. Aside from big data processing, it also ranks highest in AI performance. Riken is the national institute for research in Japan. 

Fugaku took first place as it achieved 442 petaflops, besting the United States’ Summit supercomputer built by IBM. Fugaku is a next-gen supercomputer succeeding the country’s K computer, which was the #1 in 2011. Costing $1.22 billion (130 billion yen), the system was fully operational in March 2021. Because of its high computing power, it is perfect for pharmaceutical development. With its high simulation capabilities, it is currently used to study chemicals that can be used for Covid-19 treatments. Aside from helping in the Covid-19 research, the Japan Automobile Manufacturers Association aims to use the new supercomputer to use its AI capability to study the impact of collisions, as the association wants automakers to create more resilient vehicle structures. 

Image Source:

Fugaku at a glance

Fugaku has close to 7.3 million cores and runs at a speed of 415.5 petaFLOPS (1 petaFLOP equals one thousand trillion, or one quadrillion, operations per second). With its speed, it currently outperforms Summit, whose speed is 148.6 PetaFLOPS. Fugaku brings high-performance computing (HPC) technologies a step closer to the anticipated exascale computing era. 

The supercomputer is the first among the top-ranked systems to use ARM processors. It also has two new features: a new iteration of the Tofu network and hybrid memory cubes attached to the processors. 

Fugaku is installed at the RIKEN Center for Computational Science in Kobe, Japan. The primary purpose of the supercomputer is to address high-priority scientific and social issues, such as exploring the fundamental laws of the universe, clean energy development, climate and weather forecasting, personalized medicine, and drug discovery. 

Despite not being fully operational last year, Fugaku was used in Covid-19 research.

Other top-ranked supercomputers

Summit, which debuted in June 2018, is this year’s second-ranked supercomputer after Fugaku. It is developed by IBM and installed at the Oak Ridge National Laboratory in Tennessee. The development of the supercomputer was sponsored by the U.S. Department of Energy, for the specific purpose of performing civilian scientific research. It is currently playing a critical role in Covid-19 research. 

Coming in third place is Sierra, which is located at the Lawrence Livermore National Laboratory in California. The supercomputer incorporated NVIDIA graphics processing units and IBM central processing units and was developed for the specific purpose of nuclear weapons modeling and simulations, which the U.S. National Nuclear Security Administration needs. With its 1,572,480 cores, it achieved a speed of 94.6 petaFLOPS.

China’s Sunway TaihuLight used to be at the top of the list for two years after its introduction in June 2016. It was the most powerful supercomputer in the world at that time, boasting 10,649,000 cores and a speed of 93.01 petaFLOPS. TaihuLight is installed at the National Computing Center in Wuxi, China. Scientists and engineers are using the supercomputer for marine forecasting, and various other tasks including advanced manufacturing and climate science. 

Taking the fifth spot is year is Tianhe-2 of China. It took the top spot during its debut in 2013. Tianhe-2 which means Milky Way-2 was created by the National University of Defense Technology and is intended to perform tasks for government security applications, which means that it’s involved in top-secret government projects. Tianhe-2 had several upgrades and now has 4,981,760 cores, and runs at 61.4 petaFLOPS. 

The friendly race to have the most powerful supercomputer continues. They may have different purposes, but overall, they help humans in various ways. Right now, countries with supercomputers are collaboration on critical projects, such as better weather forecasting and finding a cure for many debilitating diseases.