Japan’s Ministry of Science and Technology is planning to start the development of a supercomputer able to take 100 times the processing capacity of Japan’s current fastest machine, Fujitsu’s K computer. The working group of technological experts has finished drafting the midterm draft of the development plan, and the project is expected to be finished by 2020.
The K computer, jointly developed by the Riken research institute and Fujitsu Ltd. In 2011, can perform 10 quadrillion (10,000 trillion) calculations per second – or in computer-speak is equivalent to 10 petaflops. “Flops” stands for FLoating-point Operations Per Second, and is the computing industry’s standard measure for computer performance. The new supercomputer, K’s successor, is planned to achieve a speed of 1 exaflop, or 1 quintillion (1,000 quadrillion) calculations per second. The K computer, named after the Japanese word “kei” (10 to the 16th power), was the fastest floating-point calculating machine in the world at the time of its launch in 2011, but has since then dropped to third place last year, overtaken by IBM’s Sequoia and the Cray Titan, both from the United States.
Currently, the race to create an “exaflop” computer is joined by many European countries, the United States and even China. Japan’s science and technology ministry plans to develop an exaflop supercomputer to carry out complex simulations needed to resolve problems in social and scientific areas, like the prediction of earthquakes and the calculations involved in the subsequent evacuation processes. It will also be used to predict potential side effects of medical drugs and carry out other calculation-heavy simulations and processes.