MRI: Acquisition of a High-Performance Computing System for Research, Education, and Training

该合同将用于为伍斯特理工学院购买高性能计算（HPC）系统。目标系统由64个16核中央处理单元（cpu）和64个高性能通用图形处理单元（gpu）组成。cpu将共同提供1024个计算核心，gpu将提供每秒42.56万亿次浮点运算的总峰值性能，超过cpu。在计算能力和架构复杂性方面，该系统远远超过了WPI目前可用的任何系统。它将为主要的大规模计算任务提供急需的共享资源，为算法开发和实验提供先进的体系结构平台，并为高性能计算和应用的教育和培训提供高效的工具。它将立即用于推进研究人员的研究项目，解决精子运动和相互作用的三维建模，耦合多块多物理系统的并行解决方法，人类心室和斑块的计算建模，液冷电子设备中有效的多尺度热行为模型的计算验证，以及定点迭代的加速方法。此外，它将从一开始就提供给全校其他教师、学生和博士后研究人员使用。这种高性能计算（HPC）系统将使研究人员和其他人的研究达到新的水平，为伍斯特理工学院（WPI）的教育和培训提供新的机会，并用于提高年轻人对HPC潜力的认识，以解决社会关注的问题。研究人员将开发一门关于高性能计算方法和应用的新研究生课程，该系统将在其中发挥核心作用，他们将促进该系统在现有课程和项目中的使用，特别是在WPI的新生物信息学和计算生物学项目中。WPI杰出的基于项目的本科课程将为本科生参与HPC研究提供独特的机会，研究人员将共同开发并建议使用HPC系统的本科项目。为了进一步让本科生参与HPC应用，他们将把使用HPC系统的“真实世界”工业项目引入WPI的nsf资助的工业数学和统计本科生研究经验。为了扩大人们对高性能计算在科学和社会中的作用的认识，他们将开发演示和交互式模拟程序，使用高性能计算系统来说明如何使用高性能计算来解决社会关注的问题。已经确定了在外展活动中使用这些方案的一些可能性，并将寻求更多的可能性。此外，涉及HPC活动的实习机会将与wpi附属的马萨诸塞州数学与科学学院一起开发，这是一所面向高级学生的全州磁铁学校。

This award will be used to acquire a high-performance computing (HPC) system for Worcester Polytechnic Institute. The targeted system consists of 64 16-core central processing units (CPUs) augmented with 64 high-performance general-purpose graphical processing units (GPUs). Together, the CPUs will provide 1024 compute cores, and the GPUs will offer an aggregate peak performance rating of 42.56 trillion floating-point operations per second beyond that of the CPUs. In both computing power and architectural sophistication, this system is far beyond anything currently available at WPI. It will provide a much-needed shared resource for major large-scale computing tasks, an advanced-architecture platform for algorithm development and experimentation, and a highly effective vehicle for education and training in HPC and applications. It will be used immediately to advance research projects of the investigators that address three-dimensional modeling of sperm motility and interactions, parallel solution methods for coupled multi-block multi-physics systems, computational modeling of human ventricles and plaques, computational validation of effective multi-scale models of thermal behavior in liquid-cooled electronics, and acceleration methods for fixed-point iterations. Additionally, it will be made available from the start for use by other faculty, students, and postdoctoral researchers across the university.This high-performance computing (HPC) system will enable new levels of research for the investigators and others, provide new opportunities for education and training at Worcester Polytechnic Institute (WPI), and be used to raise awareness among young people of the potential of HPC to address problems of societal concern. The investigators will develop a new graduate course in HPC methods and applications, in which the system will play a central role, and they will promote the system's use in existing courses and programs, in particular in WPI's new Bioinformatics and Computational Biology program. WPI's distinguished project-based undergraduate program will provide unique opportunities for involving undergraduates in HPC research, and the investigators will jointly develop and advise undergraduate projects that use the HPC system. To further involve undergraduates in HPC applications, they will introduce "real world" industrial projects that use the HPC system to WPI's NSF-funded Research Experience for Undergraduates in Industrial Mathematics and Statistics. To broaden awareness of the role of HPC in science and society, they will develop programs for demonstrations and interactive simulations that will use the HPC system to illustrate how HPC can be used to address problems of societal concern. A number of possibilities for using these programs in outreach activities have been identified, and more will be sought. Also, internship opportunities involving HPC activities will be developed with the WPI-affiliated Massachusetts Academy of Mathematics and Science, a state-wide magnet school for advanced students.