The Fugaku supercomputer is located in Japan
Fujitsu and Japan’s Riken Institute of Physical and Chemical Research have announced the completion of the world’s fastest supercomputer. The Fugaku supercomputer, which has been ranked number one in the six-month Top500 ranking of the world’s most powerful supercomputers since last June , is now available for shared use by researchers around the world.
World’s fastest supercomputer completed
Development of Fugaku began in 2014. The supercomputer was created as the base system for the innovative high performance computing infrastructure (HPCI) promoted by the Japanese Ministry of Education, Culture, Sports, Science and Technology (MEXT). The deployment of all racks was completed in May 2020. Since then, partners have developed and optimized the user environment required for system sharing.
The Japan Institute of Informatics and Technology (RIST) has already selected 74 research projects that will be the first to gain access to Fugaku’s resources. However, RIST continues to accept applications from interested researchers in several categories.
The advantages of Fugaku include not only high performance, but optimization for operations often used in deep learning and artificial intelligence tasks. By combining this feature with traditional simulation, you can, for example, use AI to find the parameters needed for simulation and to extrapolate simulation results. Conversely, simulations can generate large amounts of data for learning AI.
These techniques are expected to enable Fugaku to accelerate drug development and early detection of diseases, promote preventive medicine through big medical data analysis and biological modeling, accurately predict tornadoes and heavy rainfalls, simulate earthquakes, tsunamis and escape routes, develop low-cost and highly efficient solar cells, fuel cells and storage batteries, accelerate the creation of new devices and materials, form a deeper understanding of the fundamental issues of science.
As a reminder, Fugaku uses Fujitsu A64FX SoCs based on Arm architecture. A system with 7 630 848 cores entered the November Top500 list with a result of 442 petaflops in the High Performance Linpack (HPL) test.
