MRI: Acquisition of a GPU-Accelerated Research Cluster

大规模数据分析能力已成为推动科学各领域快速进步的关键因素。该项目将通过收购支持大数据驱动研究的最先进的GPU计算机器，显著扩大布兰代斯大学的高性能计算资源。新的能力将支持跨不同学科的综合研究项目，包括生物学、生物化学、化学、计算机科学、数学、物理和心理学领域。它还将为布兰代斯大学提供一个平台，培训下一代劳动力开发和应用深度学习技术，从而加速基础研究和技术创新的发现。该项目将使Brandeis的研究人员和教师能够利用现代计算技术扩大对科学、技术、工程和数学(STEM)领域的参与。具体地说，新资源将包括10 Gbit网络结构上的16个GPU节点和1个存储节点，以实现快速节点间通信。该项目将使Brandeis的研究人员能够进行大数据驱动的融合研究，提高我们对从神经科学到病毒组装等领域的生命规则的理解，并解决潜在的社会需求(如绿色能源)的根本问题。新兴的跨学科研究活动还将为开发新的、更强大的大数据驱动技术创造一个包容的环境。该项目将使培训教师、博士后、研究生和本科生有效使用最先进的GPU计算和深度学习技术的课程和研讨会成为可能。它将允许Brandeis通过目前由NSF资助的位于Brandeis材料研究科学和工程研究中心的REU站点，进一步整合其教育和研究。此外，它还将通过几个现有的项目，包括NSF资助的REU站点、针对经济困难学生的过渡年计划、当地的芝加哥人/拉美裔美国人和印第安人在Brandeis的科学分会、Brandeis科学Posse计划和Brandeis课堂工作坊的Brandeis科学家研讨会，加强Brandeis扩大STEM参与的能力。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Large scale data analysis capability has become the key factor that is driving rapid progress in all fields of science. This project will significantly expand the high-performance computing resources at Brandeis University, by acquiring state of the art GPU computing machines that enable big data driven research. The new capability will support integrated research projects across a variety of disciplines, including the areas of Biology, Biochemistry, Chemistry, Computer Science, Math, Physics, and Psychology. It will also provide a platform at Brandeis University for training the next generation workforce to develop and apply deep learning techniques, thus accelerating discoveries in basic research and technological innovations. The project will enable Brandeis researchers and instructors to utilize modern computing techniques to broaden participation in science, technology, engineering, and mathematics (STEM) fields. Specifically, the new resources will include 16 GPU nodes and 1 storage node on a 10Gbit network fabric to allow rapid internode communications. This project will enable Brandeis researchers to conduct big data driven convergence research that enhances our understanding of the Rules of Life in areas ranging from neuroscience to virus assembly, and addresses the fundamental problems underlying societal needs (e.g., green energy). The emerging interdisciplinary research activities will also create an inclusive environment for developing novel and more powerful big data driven techniques. The project will enable courses and workshops that train faculty, postdocs, graduate students, and undergraduate students to effectively use state of the art GPU computing and deep learning techniques. It will allow Brandeis to further integrate its education and research via a current NSF funded REU site at Brandeis Materials Research Science and Engineering Research Center. In addition, it will enhance Brandeis' ability to broaden participation in STEM, especially by women and underrepresented minorities, through several existing programs, including the NSF funded REU site, the Transitional Year Program for economically disadvantaged students, the local Society for Advancement of Chicanos/Hispanics and Native Americans in Science chapter at Brandeis, the Brandeis Science Posse Program, and the Brandeis Scientists in the Classroom Workshop.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.