Focused on promoting scientific discovery through the application of advanced computing technologies, CCS supports computational scientific research in Japanese universities and institutes by operating advanced computing systems. The Center performs research on critical issues in fundamental science, material science and environmental science by performing large-scale simulations and large-scale data analysis.
Utilizing Intel Xeon processors and NVIDIA Tesla K20X GPU accelerators, the new Cray CS300 cluster supercomputer is custom engineered for Tsukuba CCS by integrating its Tightly Coupled parallel computing Acceleration (TCA) mechanism into a standard Cray CS300 system. This new system has been combined with the University's existing Cray cluster, which is the Highly Accelerated Parallel Advanced system for Computational Science (HA-PACS).
"We are pleased to introduce our new Cray CS300 cluster supercomputer with the latest NVIDIA Tesla K20X GPUs as an extension of our GPU cluster HA-PACS", stated Prof. Taisuke Boku, a deputy director and system manager at CCS. "The combined HA-PACS/TCA system is equipped with our hardware technology that enables GPU to GPU direct communication over nodes. We can now open a new frontier of accelerated, parallel computing for a wide variety of computational sciences."
"The University's Center for Computation Sciences is one of the premier high performance computing facilities in Japan, and we are honored the Center once again turned to Cray to provide its users with the computational resources of a Cray CS300 cluster supercomputer", stated Mamoru Nakano, president of Cray Japan. "The Center is also widely recognized in the Japanese supercomputing community as a leader in accelerated computing, and we are pleased that the Cray CS300 system paired with GPU accelerators provided the right price/performance solution to meet the needs of their applications."
The Cray CS300 series of cluster supercomputers are scalable, cluster solutions that group optimized, industry-standard building block server platforms into a unified, fully-integrated system. Available with air or liquid-cooled architectures, Cray CS300 systems provide superior price/performance, energy efficiency and configuration flexibility. The systems are integrated with Cray's HPC software stack and include software tools compatible with most open source and commercial compilers, schedulers, and libraries. Cray CS300 systems also feature the Cray Advanced Cluster Engine, an essential management software suite designed to provide network, server, cluster and storage management capabilities that are necessary to run large, complex technical applications.