NVIDIA GPUs Power Fujitsu AI Supercomputer at RIKEN in Japan

Fujitsu has posted news that their new AI supercomputer at RIKEN in Japan is already being used for AI research. Called RAIDEN (Riken AIp Deep learning ENvironment), the GPU-accelerated Fujitsu system sports 4 Petaflops of processing power.

The RAIDEN supercomputer is built around Fujitsu PRIMERGY RX 2530 M2 servers with and 24 NVIDIA DGX-1 systems. With 8 NVIDIA Tesla GPUs per chassis, the DGX-1 includes access to today’s most popular deep learning frameworks, NVIDIA DIGITS deep learning training application, third-party accelerated solutions, the NVIDIA Deep Learning SDK (e.g. cuDNN, cuBLAS, NCCL), CUDA toolkit, NVIDIA Docker and NVIDIA drivers.

Artificial intelligence is expected to be a breakthrough in technological leadership worldwide. Because Japan was considered to be lagging behind in this field, in 2016 the Ministry of Education, Culture, Sports, Science and Technology launched the the AIP (Advanced Integrated Intelligence Platform).

In the past, individual researchers had been separately conducting research, but then it is difficult to make a big leap forward,” said AIP center director Masashi Sugiyama. “When the country has a sufficient budget and it gathers researchers from various fields, multiple research can be combined. I hope that the researchers will be inspired each other and other effects.”

RAIDEN Supercomputer Specifications

According to Sugiyama, the AIP center will be and international effort. “I want to foster this as an international research base. It is not enough for Japanese researchers alone for AI research. Is internationalization of researchers can not be said to be an essential condition for center success.”

Fujitsu has systematized knowledge and technology on AI, which it calls “Human Centric AI Zinrai” and actively engages in AI field. Officials at RIKEN are looking forward to continued cooperative research.

We are working on fundamental research, but there is a strong desire to utilize it in the real world. I want to bring algorithms to practical use and bring them to the point of spreading to the world,” Mr. Sugiyama said.

Sign up for our insideHPC Newsletter