The National Institute of Advanced Industrial Science and Technology (AIST) in Tokyo has ordered a supercomputer system from Fujitsu as it looks to build a cloud-based "open innovation platform" for artificial intelligence applications called AI Bridging Cloud Infrastructure (ABCI).
Fujitsu's supercomputer system will offer a theoretical peak performance in half-precision floating point operations of 550 petaflops, and 37 petaflops of double-precision floating point operation performance, the publicly-listed Japanese IT giant has said.
The system will consist of 1,088 Primergy CX2570 M4 servers, mounted in Fujitsu's Primergy CX400 M4 multi-node servers, with each server featuring components such as two Intel Xeon Gold processor CPUs, four NVIDIA Tesla V100 GPU computing cards, and Intel SSD DC P4600 storage.
Fujitsu said it will construct the system with the goal of commencing operations in April 2018.
Japan's Ministry of Economy, Trade and Industry said last year that it would spend 19.5 billion yen on the ABCI platform in a bid to propel the nation to the top of the supercomputing race.
The platform is aimed at enabling high-speed AI processing, and subsequently accelerating the deployment of artificial intelligence into "real businesses and society", the AIST has previously said.
ABCI will be deployed to the new ABCI datacentre on the Kashiwa II campus of the University of Tokyo.
Earlier this year, the Research Institute for Information Technology at Kyushu University in Fukuoka, Japan announced it will be receiving a new supercomputer system from Fujitsu this month, which will be used by national universities to advance research in areas such as artificial intelligence.
The university said at the time that it will use the new system as a computational resource for the JHPCN Joint Usage/Research Center for Interdisciplinary Large-scale Information Infrastructures, which is a network of joint use locations made up of supercomputer facilities at numerous universities across Japan, with the Information Technology Center at the University of Tokyo serving as the core location.
The back-end system will be made up of 2,128 Primergy CX400 systems and Intel Skylake Xeon processors, with 433 terabytes of total memory capacity. 128 of the x86 servers will be equipped with four Nvidia Tesla P100 GPU computing cards.
Front-end subsystem will comprise 160 basic front-end nodes featuring Intel Skylake Xeon processors and Nvidia Quadro P4000 graphics cards, as well as four high-capacity front-end nodes featuring 12 terabytes of memory each, in addition to other servers.
Along with a theoretical peak performance of 10 petaflops, the system will offer a 24 petabyte storage system, 100Gbps InfiniBand EDR interconnect, and Fujitsu's scalable cluster file system, FEFS.
The Riken Center for Advanced Intelligence Project in Japan received its own deep learning supercomputer in April, which will be used to accelerate research and development in the "real-world" application of AI.
The system is comprised of two server architectures, with 24 Nvidia DGX-1 servers -- each including eight of the latest Nvidia Tesla P100 accelerators and integrated deep learning software -- and 32 Fujitsu Server Primergy RX2530 M2 servers, along with a high-performance storage system.
Its file system is also FEFS on six Fujitsu Server Primergy RX2540 M2 PC servers; eight Fujitsu Storage Eternus DX200 S3 storage systems; and one Fujitsu Storage Eternus DX100 S3 storage system to provide the IO processing demanded by deep learning analysis.
Fujitsu reported consolidated revenues of 4.5 trillion yen for the fiscal year ended March 31, 2017.