Case Study: Osaka University achieves faster AI and HPC research with Supermicro SuperBlade and GPU systems

Osaka University Creates AI Supercomputer with Supermicro SuperBlades Utilizing 3rd Gen Intel Xeon Scalable Processors

Osaka University required a significant upgrade to its supercomputing infrastructure to meet the increasing demands of its diverse research community. The university needed faster processing, a reduced time to solution for complex simulations in fields like AI, HPC, and data analysis, and a more balanced and efficient system with a lower total cost of ownership and carbon footprint. This challenge led them to partner with vendor Supermicro for a new solution.

Supermicro implemented a comprehensive supercomputing solution named SQUID, consisting of 27 racks of its SuperBlade systems and 4U 8 GPU servers, all powered by 3rd Gen Intel Xeon Scalable Processors and featuring direct liquid cooling. This deployment delivered substantially faster simulation times, enabled higher-resolution research models, and dramatically accelerated AI, machine learning, and deep learning workloads by efficiently leveraging the latest GPUs. The result allows Osaka University researchers to run applications across over 120,000 cores simultaneously, significantly accelerating the pace of scientific discovery.

Osaka University

Susumu Date

Cybermedia Center