January 8, 2014

Japan Moves Forward with 2020 Exascale Plans

Tiffany Trader
Japan HPCI graphic

In the race to build the first exascale supercomputer, several nations have set their sites on 2020 as the make-or-break year. With six years to go until that deadline, we can expect exascale efforts to take on a new urgency. Japan, which is among the economic regions seeking exascale glory, just got a little closer to that goal. On December 26, the Japanese Ministry of Education, Culture, Sports, Science and Technology selected RIKEN to develop an exascale supercomputer by 2020.

RIKEN, the large Japanese research institution which relies on supercomputing to pursue advances in a diverse range of scientific disciplines, has been tasked with keeping Japan at the leading edge of computing science and technology.

“Exascale supercomputing is expected to make possible high-resolution simulations, contributing to advances in a wide range of areas including drug discovery, weather forecasting, and astrophysics,” noted the Ministry in a December 26th statement.

RIKEN was selected for this seminal project based on its experience developing and operating Japan’s current top number cruncher, the K computer, which was ranked as the fastest supercomputer in the world in 2011. With 10.51 petaflops LINPACK, the system is still a strong TOP500 contender, as the current number-four machine. Getting to exascale will require fielding a system that is 100 times faster than “K” and 30 times faster than the current TOP500 record-holder, China’s Tianhe-2 supercomputer, which achieved 33.86 petaflops on the LINPACK benchmark.

“The RIKEN Advanced Institute for Computational Science (AICS) will now have two important missions,” notes Kimihiko Hirao, director of the RIKEN Advanced Institute for Computational Science (AICS) of Japan, “continuing to operate and manage the K computer for public use with the aim to generate useful research outcomes, and the successful development of the Exascale Supercomputer scheduled for completion by 2020. We ask for support from our associates around the world and in Japan as we launch our new project, which will be a great boon for science and technology, as well as industry.”

While nations like the United States, China, Japan and the EU would like to be first to break the exascale barrier, advancing science and technology also requires collaboration, sharing and a degree of openness. At SC13, the US and Japan signed a memorandum of understanding (MOU) in recognition of a new partnership focused on expanding the use of petascale computing in the scientific and engineering communities. Aside from demonstrating that there is still much petascale-level work to be done, the initiative reflects the importance of global collaboration in the pursuit of scientific progress. Going forward, it will be interesting to see how exascale-focused nations balance the drive to compete with the advantages of cooperation.

Tags: , ,