CC* Compute: Accelerating Science and Education by Campus and Grid Computing

点击翻译按钮获取中文摘要

High-performance computing resources are a fundamental need of modern research that unites almost all disciplines. Experience with these resources is also an important tool for preparing today's students for a wide range of careers. A group of researchers and educators at the University of Colorado Denver, in partnership with its Office of Information Technology, are building a state-of-the-art computing resource on its Downtown Campus. The new facility provides the first campus-wide high-performance computer system to support both research and teaching efforts. The computing cluster is integrated with the Open Science Grid (OSG), enabling access to additional computing resources from partner institutions, and sharing unused time with the wider community. A high-priority educational queue is dedicated for teaching and course-based research. The resource will include 2048 AMD EPYC compute cores and 16TB memory distributed across 32 compute nodes; 2 high-memory nodes, each with 2TB memory and 64 cores; one NVIDIA Tesla V100 32GB GPU; 1PB (raw) storage; and InfiniBand interconnect. For data-intensive research and access of OSG jobs to distributed data, the cluster is configured with full end-to-end 10gb/s connectivity from each node to Internet 2. A graphical Jupyter notebook interface increases accessibility. The configuration addresses computing requirements based on a survey and an analysis of the needs of science and educational drivers in fields including earth and environmental sciences, biotechnology and genomics, computer science and engineering, applied mathematics, physics, and business. The resource will broaden participation in computational science and have a significant impact on the supported research.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.