Japan is Aiming to Build the Fastest Supercomputer in the World
Japan has allocated $173 million for the project.
Beating the Fastest
In the the world of computing, speed is king. Companies and countries spend millions of dollars just to achieve the distinction that goes with having the fastest computer in the world; a position that China has dominated for three years thanks to its two supercomputers, the Tianhe-2 and the Sunway TaihuLight.
Japan is now bidding for a place back into the supercomputer rankings by planning to build a new supercomputer that would beat the current record holder, the Sunway TaihuLight.
The Ministry of Economy, Trade and Industry will spend around $173 million on the project according to a budget breakdown reported by Reuters. The project could see its completion as early as next year.
The machine will be capable of running 130 quadrillion calculations per second or 130 petaflops which, at that speed, beats Sunway TaihuLight’s speed of 93 petaflops. Japan’s current fastest supercomputer is Fujitsu’s Oakforest-PACS which is capable of 13.6 petaflops.
The computer will be built in Japan’s National Institute of Advanced Industrial Science and Technology. Satoshi Sekiguchi, a director general of the institute tells Reuters; “As far as we know, there is nothing out there that is as fast.”
Supercomputers for AI
The new supercomputer is dubbed AI Bridging Cloud Infrastructure or ABCI and the bidding for this project has started and it will end on December 6 of this year.
Japan aims to use this supercomputer for the development of new AI technologies such as deep learning. Eventually, Sekiguchi says, Japan’s supercomputer could be used in the medical field where new treatments could be developed by pulling out medical records.
Supercomputers are useful for calculating huge amounts of data and current uses include weather forecasting, pharmaceutical development, and scientific research.
Bumping up the speed of supercomputers is sure to boost the capabilities of newer technologies like the development of better driverless cars which rely on the real-time analysis of traffic data.