MRI: Acquisition of Cutting-Edge GPU and Phi Nodes for the Interdisciplinary UMBC High Performance Computing Facility

点击翻译按钮获取中文摘要

This project will expand the interdisciplinary University of Maryland Baltimore County (UMBC) High Performance Computing Facility (HPCF), the community-based, interdisciplinary core facility for scientific computing and research on parallel algorithms at UMBC. The expansion will support the research projects of 51 researchers from 13 academic departments and research centers across the entire campus, including the areas of Computer Science, Information Systems, Mathematics, Statistics, Physics, Biology, Chemistry, Marine Biotechnology, Environmental Systems, Engineering (Computer, Electrical, Mechanical, Chemical, and Environmental), and research centers focused on environmental research, earth sciences, and imaging research.Specifically, the expanded computational facility will comprise a total of 84 compute nodes including cutting-edge NVIDIA GPU accelerators and Intel Xeon Phi KNL processors. The availability of the new resource will give researchers at UMBC the opportunity to increase scientific discovery significantly through the dramatic speedup in their simulation and modeling activities from state-of-the-art CPUs and cutting-edge GPUs and Phi KNL processors. An existing cluster at HPCF has already attracted a broad user base through a winning combination of sufficient hardware, tight integration of student education, freely available user support, and an appropriate usage policy. Moreover, the new expanded resources of HPCF will enabled UMBC to develop a powerful synergy between research and education at all levels. Through the project's consulting approach to user support, application researchers and their post-docs, graduate students, and undergraduate students will be exposed to the power of state-of-the-art computing software and hardware, a crucial experience for the future workforce. Synergistic integration of education and research is concretely exemplified by current NSF-funded initiatives at UMBC, including an REU Site on high performance computing, a proposed REU Site in quantitative biology, proposed CyberTraining initiatives, and a growing number of courses that use HPCF. HPCF also actively partners with other efforts on campus, such as the UMBC Meyerhoff Scholarship and the NIH-funded MARC programs, two nationally recognized programs that attract substantial numbers of students from underrepresented groups into the sciences.