Please view the main text area of the page by skipping the main menu.

Editorial: Hopes for Japanese supercomputing to reach new heights

Japan's K supercomputer is seen at the Riken Center for Computational Science, formerly the Riken Advanced Institute for Computational Science, in Kobe's Chuo Ward on Jan. 7, 2016. (Mainichi/Kentaro Ikushima)

The government-backed Riken research institute wound down the use of its 'K' supercomputer in Kobe in western Japan on Aug. 16. Since September 2012 the supercomputer had served as the country's flagship machine, supporting fundamental research and companies' technological development. The computer had a utilization rate of 94% and it contributed to more than 1,300 research papers.

Supercomputers are able to perform an immense number of calculations in an instant. They are able to simulate complex phenomena on a global scale that can't be produced in laboratories, and can improve the efficiency of experiments that involve a large number of operations.

Projects utilizing K's power made contributions to safety and security in Japan. These included methods to accurately predict localized downpours and the formation of typhoons, damage predictions for buildings in the heart of Tokyo in the event of major earthquakes, and the development of anticancer drugs and fuel-efficient tires.

K's operational life of seven years is longer than the norm in the world of supercomputers, which are typically used for four to five years. Supercomputing calculation speeds are said to be increasing at a rate of about 1.8 times per year, and K had fallen from the top spot in the world in terms of speed to 20th. However, under another index measuring the computer's ability to perform complex calculations, it had held the top spot from 2015.

Some 111.1 billion yen in public funds was spent on the development of K. In the autumn of 2009, when the supercomputer was being built, officials reviewing public projects under the former Democratic Party of Japan-led administration sarcastically questioned the huge amount being invested in the project, with one lawmaker asking, "What's wrong with being No. 2?" Since it takes time for research to be put to practical use, it is not easy to measure the investment effect of K, but taking cost reductions into consideration, one U.S. firm has calculated that it would be in the range of about 1 trillion yen.

The development cost of the K computer's successor, Fugaku, is said to be about 130 billion yen, of which 110 billion yen will come from state coffers. Officials aim to have it up and running in 2021 at the same site as K, following the removal of the older supercomputer. Until then, they will make do with supercomputers at 11 university and research facilities in Japan.

Fugaku is supposed to have 100 times the computing power of K. By incorporating a design concept, like that seen in smartphones and other devices, developers hope to improve the user-friendliness of the supercomputer. It is hoped that application of the computer will widen in an age of data utilization, such as by analyzing big data and using artificial intelligence to make predictions based on the results.

International competition in the field is intense, including from supercomputer power the United States, and China, which in recent years has been investing a large amount of military funds into supercomputing. Japan cannot match these countries in the scale of investment, but it can build on the accomplishments of K, and instead of vying for the No. 1 position, underscore the new computer's presence with its overall computing power and peaceful use.

The name Fugaku comes from an alias for Mount Fuji, and we hope that like the mountain in central Japan, the new supercomputer will have a high peak and a wide base, expanding new possibilities.

Also in The Mainichi

The Mainichi on social media