MRI: Acquisition of a High-Performance Computing Cluster for Science and Engineering Research at the University of Kansas

将在堪萨斯大学(KU)建立一个高性能计算(HPC)集群资源，以促进堪萨斯大学和选定的主要本科机构(PUI)的计算科学和工程。这个绰号为BigJay的集群将成为KU社区集群的一部分，该集群位于最先进的高级计算设施中，由KU研究计算中心(CRC)维护。主要用户来自科罗拉多大学11个不同的学术部门/项目：航空航天工程、生物工程、化学、化学和石油工程、计算生物学、电气工程和计算机科学、地理学、地质学、数学、分子生物科学和物理与天文学。此外，来自四个与科罗拉多州立大学有联系的主要本科院校(PUI)的教职员工和本科生研究人员将可以访问启用的计算资源。此外，儿童基金会和参加者将举办有关集群使用和先进计算方法的工作坊，这些工作坊也将提供给来自昆士兰大学和参与的PUI的学生(和教员)。拟议的高性能计算资源将使研究生和高级本科课程的学生接触到高性能计算技术。最后，BigJay集群将促进以HPC为中心的跨学科合作的增加，包括与PUI教师参与者的合作。BigJay将由CPU、高内存CPU和支持GPU的处理器混合组成，并包括千兆以太网和高速Infiniband网络。选择拟议的配置是为了反映参与者的不同计算需求。由BigJay实现的基础研究将跨越算法开发、计算方法、机器学习和物理系统的理论描述。这项工作将影响到对社会具有重要意义的问题，从电池材料的发展到癌症识别到催化系统的优化，以及改进的物理和生物系统理论。跨学科协作将通过研讨会促进，这些研讨会将介绍跨学科的学生和教职员工。这些将被用作确定研究主题、数值方法和计算工具的共性的机会。此外，还将举办一个由海报和简短演讲组成的计算研究研讨会，每两年为与会者以及在密歇根大学校园从事计算研究的其他人举行一次研讨会，以鼓励合作。拟议的收购将大大加强KU和PUI的HPC研究工作，并使参与的研究小组在联邦、非营利性或行业来源的资金方面具有潜在的更强竞争力。它将进一步用于通过研究、CRC和参与者主导的研讨会以及将HPC组件纳入研究生和高级本科课程来加强学生在科学和技术计算方面的培训。该项目由主要研究仪器(MRI)计划、既定的激励竞争研究计划(EPSCoR)和计算机与信息科学与工程(CESE)理事会共同资助。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。

A High Performance Computing (HPC) cluster resource will be built at the University of Kansas (KU) that will advance computational science and engineering at KU and selected Primarily Undergraduate Institutions (PUIs). This cluster, nicknamed BigJay, will be part of the KU Community Cluster housed in the state-of-the-art Advanced Computing Facility and maintained by the KU Center for Research Computing (CRC). The major users form a multidisciplinary group from 11 different academic departments/programs at KU: Aerospace Engineering, Bioengineering, Chemistry, Chemical and Petroleum Engineering, Computational Biology, Electrical Engineering and Computer Science, Geography, Geology, Mathematics, Molecular Biosciences, and Physics & Astronomy. In addition, faculty and undergraduate researchers from four different Primarily Undergraduate Institutions (PUIs) with ties to KU (The College of St. Scholastica, Prairie View A&M University, Hobart & William Smith College, and the University of Southern Indiana) will have access to the enabled computational resources. Additionally, workshops will be offered by the CRC and participants on cluster use and advanced computational methods, that will also be available to students (and faculty) from KU and the participating PUIs. The proposed HPC resources will enable exposure to HPC techniques to students in graduate and advanced undergraduate courses. Finally, the BigJay cluster will stimulate increased interdisciplinary collaborations centered around HPC, including with the PUI faculty participants. BigJay will be made up of a mix of CPU, high-memory CPU, and GPU-enabled processors and include gigabit ethernet and high-speed Infiniband networking. The proposed configuration was chosen to reflect the diverse computing need of the participants. The fundamental research enabled by BigJay will span algorithm development, computational methodology, machine learning, and theoretical descriptions of physical systems. The work will impact problems of importance to society ranging from the development of battery materials to cancer recognition to optimization of catalytic systems, as well as improved theories of physical and biological systems. Interdisciplinary collaborations will be facilitated by the workshops that will introduce students and faculty across disciplines. These will be used as an opportunity to identify commonalities in research topics, numerical methods, and computational tools. Collaborations will also be encouraged by instituting a Computational Research Symposium, consisting of posters and short talks, that will be held biannually for the participants, as well as others who are engaged in computational research on the KU campus. The proposed purchase will substantially enhance HPC research efforts at KU and the PUIs and make participating research groups potentially more competitive for funding from federal, non-profit, or industrial sources. It will further be used to enhance student training in scientific and technical computing through research, the CRC and participant-led workshops, and inclusion of HPC components in graduate and advanced undergraduate courses.This project is jointly funded by the Major Research Instrumentation (MRI) program, the Established Program to Stimulate Competitive Research (EPSCoR), and the Computer & Information Science & Engineering (CISE) Directorate.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.