Japan’s National Institute of Advanced Industrial Science and Technology (AIST) has selected HPE to build its next-generation supercomputer, ABCI (AI Bridging Cloud Infrastructure) 3.0. This supercomputer will be offered as a cloud service to public and private organizations, supporting large foundational models for generative AI to accelerate research, development, innovation, and social implementation.
AIST’s new ABCI 3.0 will be constructed using HPE Cray XD systems featuring NVIDIA H200 Tensor Core GPUs, interconnected by NVIDIA Quantum-2 InfiniBand networking. The supercomputer is expected to achieve a theoretical maximum of approximately 6.2 exaflops of half-precision (16-bit) performance.
Slated to be the fastest AI supercomputer in Japan, ABCI 3.0 will help meet the high demand for the country’s growing number of AI workloads. Since 2018, AIST has made its cloud supercomputing services accessible to research institutions, universities, and commercial companies of all sizes, from startups to major electronics manufacturers. AIST’s ABCI systems have supported a range of scientific research, including the development of AI techniques for genome sequencing prediction, materials informatics, and Japanese large language models (LLMs).
