MRI: Acquisition of SeaWulf - A Reconfigurable Computer System for Research and Education

该项目部署了一个高性能计算机系统，以加强石溪大学（SBU）和纽约州合作机构的研究、教育和培训活动。研究课题包括模拟风暴潮和相关损害，为弹性沿海基础设施和应急响应的设计提供信息；基因组学分析探索生物多样性和如何保护未来的生态系统；分析计算机系统内的数据运动，以实现更节能的设计；发展可扩展的脑功能模型；拟用于聚变反应堆的材料的原子模拟；超新星和x射线爆发的模拟是天文学的基础。SBU的60多个研究生和本科课程经常使用该系统，这也为本科生的研究创造了许多机会。通过与SBU全纳教育中心（CIE）和其他组织的合作，该工具的运作大大扩大了妇女、代表性不足的少数民族、残疾人和高中生对科学和工程研究的参与。外部合作伙伴包括纽约州立大学系统和新罗谢尔学院的女子艺术与科学学院。该仪器通过命令行和门户网站访问支持高级和入门级用户，并且可重新配置以在批处理和云工作流程之间移动资源。它包括165个节点，通过InfiniBand连接，共有4516个内核，总内存为20.7TB，并由一个快速的1PB通用并行文件系统（GPFS）支持。高性价比、高性能和高生产率的计算加速和扩展了研究，并为研究项目和用户利用更大的国家计算设施做好了准备。一旦投入使用，该系统将用于SBU的40个nsf资助的研究项目，以及由工业或其他机构支持的多个其他项目，包括空军科学研究办公室（AFOSR）、美国国家航空航天局（NASA）、能源部（DOE）和纽约州。

The project deploys a high-performance computer system that enhances the research, education and training activities at Stony Brook University (SBU) and partner institutions across the state of New York. Research topics include modeling storm surges and associated damages to inform the design of resilient coastal infrastructure and emergency response; genomics analysis to explore biodiversity and how to conserve ecosystems into the future; analyzing data motion within computer systems to enable more energy efficient designs; development of scalable models of brain function; atomistic simulations of materials proposed for use in fusion reactors; and simulation of supernovae and x-ray bursts fundamental to astronomy. Over sixty graduate and undergraduate courses at SBU regularly use the system, which also creates numerous opportunities for undergraduate research. Through collaborations with the SBU Center for Inclusive Education (CIE) and other organizations, the instrument operates to substantially broaden participation in science and engineering research by women, underrepresented minorities, persons with disabilities, and high-school students. External partners include the SUNY system and the all-women School of Arts & Sciences at the College of New Rochelle. The instrument supports both sophisticated and entry-level users through command-line and web-portal access, and is reconfigurable to move resources between batch and cloud workflows. It includes 165 nodes connected via InfiniBand for a total of 4516 cores with 20.7TB total memory, and is supported by a fast 1PB General Parallel File System (GPFS). The cost-effective, high-performance and high-productivity computing accelerates and expands research, and prepares the research projects and users to exploit the larger national computing facilities. Once operational the system will be used for research and training by forty NSF-sponsored research projects at SBU, and by multiple other projects supported by industry or other agencies, including the Air Force Office of Scientific Research (AFOSR), the National Aeronautics and Space Administration (NASA), the Department of Energy (DOE), and the State of New York.