Fujitsu has announced that it has received an order from RIKEN, Japan’s largest research institution, for the development of “Deep Learning System”. In terms of operations, this is going to be one of the largest-scale supercomputers in Japan that would be dedicated to AI. The new system is scheduled to go online in April where it would be utilised as a platform for accelerating R&D in the field of Artificial Intelligence. 

 

Apart from a high-reliability and storage facility, the system will mainly comprise two server architectures including 24 NVIDIA DGX-1 servers and 32 PRIMERGY RX2530 M2 servers. 

 

In order to build one of Japan’s most advanced AI research systems, the company is leveraging its extensive knowledge regarding AI and computing. Fujitsu would also extend support for R&D, thereby paving the path for a society that would use AI to solve various social issues. 

 

 

 

The system is built of two server architectures while making use of the latest CPUs and GPUs alongside a storage system. This system is being installed at Fujitsu’s Yokohama data centre. The company has integrated NVIDIA’s DGX-1 deep learning software environment with a customized software environment to be used on a secure on-site network. 

 

The system utilizes operations management functions for enabling creation of calculation executive environments. It’s also highly secure and reliable for processing data with high importance like personal and intellectual property data.

 

 

The configuration of the Deep Learning System can be explained as follows:

 

 

It contains 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. Besides, it has a total theoretical performance of more than 4 *Petaflops. 

 

 

6 FUJITSU Server PRIMERGY RX2540 M2 PC servers, 8 FUJITSU Storage ETERNUS DX200 S3 storage systems, and 1 FUJITSU Storage ETERNUS DX100 S3 storage system are available to provide the IO processing demanded by deep learning analysis.

 

As per, Mr. Jim McHugh, VP and General Manager at NVIDIA, “NVIDIA DGX-1, the world’s first all-in-one AI supercomputer, is designed to meet the enormous computational needs of AI researchers. Powered by 24 DGX-1s, the RIKEN Center for Advanced Intelligence Project’s system will be the most powerful DGX-1 customer installation in the world. Its breakthrough performance will dramatically speed up deep learning research in Japan, and become a platform for solving complex problems in healthcare, manufacturing and public safety.”

 

*Petaflops: Petaflops is short for peta floating point operations per second, where Peta is an SI prefix indicating one quadrillion, or 1015. This denotes one quadrillion floating point operations per second.