MRI: Acquisition of a GPU-Accelerated Parallel Supercomputer for Computational Science and Engineering Research at the University of New Mexico

点击翻译按钮获取中文摘要

We will acquire a graphical processing unit (GPU)-based 32-node parallel high-performance computer in support of science and engineering research at the University of New Mexico (UNM). The machine will be housed at the UNM Center for Advanced Research Computing (CARC), a campus-wide shared facility. The machine will significantly expand the capabilities of the campus supercomputer center (CARC), increasing total available compute cycles from 5 to 41 TFlops peak, and corresponding online storage capacity from 23 to 55 TB. Configured as a tightly-coupled cluster of "fat nodes", the new machine and associated software will enable researchers to run calculations as intrinsically parallel (multi-threaded, 10-50x GPU-accelerated) single-node, shared-memory jobs, with further scalability achievable through multi-node message-passing parallelism.The seventeen collaborating faculty on this proposal span nine departments and three Colleges. This new system will support research in nano-bio-materials science, including molecular biophysics, chemical and condensed matter physics, materials physics, mathematical biology, molecular biology, catalysis, novel sensor materials, and structural materials; advanced graphics, image processing and visualization, including biophysical imaging using quantum dots, 3D animation and rendering, fMRI image analysis and 3D computed tomography; and geophysics, including computational electromagnetics and geological modeling. It will also provide the opportunity for students and researchers to remain on the computing technology curve and gain early access to a transformative next- generation architecture. Outreach and training classes will be provided by CARC technical staff to significantly expand the user base, particularly to 'nano-bio' researchers (students and faculty) associated with the two NSF IGERT training programs at UNM, Nanoscience and Microsystems and Integrating Nanotechnology with Cell Biology and Neuroscience, as well as the Initiatives to Maximize Student Diversity (IMSD) program. IMSD is aimed at increasing the number of under-represented minorities in biomedical research and has students from Biology, Biochemistry, Chemistry, Computer Science, Computer Engineering, Chemical Engineering and Psychology. Students and researchers from the MIND Research Network (neuroscience) and the NSF-funded Data Observation Network for Earth (DataOne) project (environmental sciences) will also have access to and benefit from the machine. As an EPSCoR and Minority-Serving Institution, UNM has a unique demographic makeup: it is the only Carnegie Research/Doctoral Extensive institution in the U.S. that is also a Hispanic Serving Institution, with 32% Hispanic, 6% Native American and 3% African American students. The availability of this next-generation supercomputing architecture as a shared, campus-wide resource will provide an exceptional opportunity to engage students, including many from traditionally underrepresented minorities, at the forefront of computational science and engineering research.