Acquisition of a Beowulf Cluster for Computational Materials Research, Education and Student Training

这笔赠款为收购Beowulf集群提供支持，用于密歇根大学的计算材料研究、教育和学生培训。由于计算机能力和可获得性的惊人进步，模拟现在指导着纳米科学和技术、微电子学、聚合物加工、光子学、设计材料、自主材料系统和许多其他重要技术领域的材料过程的发现和理解。解决材料问题所需的模拟技术本质上是多学科的，涉及跨越许多数量级的长度和时间尺度。因此，它们需要组合使用最初为特定学科中的特定应用构思的不同的计算方法。这些不同技术的专业知识往往在其起源领域内仍然是孤立的。必须消除这些学科界限，以便开发协调和整合互补计算方法的跨尺度模拟方法，从而促进对过于复杂而不能用单一技术解决的问题的调查。其目标有三个：(1)促进密歇根大学在计算材料方面的跨学科和创新研究和教育；(2)领导材料研究的尖端、跨越规模的模拟方法的开发和应用；以及(3)教育和培训学生，使他们掌握从分子到材料发现、设计和开发的从分子到宏观性质的计算方法。我们预计我们的毕业生将被行业、政府实验室和大学大力招聘。由于需要汇集不同计算方法的专业知识，因此这项工作必须建立一个中心形式。通过计划的各种项目，该中心将为教师和学生之间的互动提供一个论坛，以促进推动计算材料研究进步的许多专业领域之间的交叉培养。该集群将允许多个用户同时用于教育、培训和研究，并将允许对材料和材料过程的模拟，在某些情况下比以前可以达到的数量级更大。这种规模上的改进对于促进基本理解和科学发现至关重要。该项目的更广泛影响是通过开发创新的计算方法和培训高技能的计算科学家和工程师来实现的，能够满足当今社会对材料技术进步日益增长的需求。我们的愿景是在计算材料科学和工程的研究和教育方面成为引人注目的领导者，在密歇根开发模拟技术、课程和推广计划，作为世界各地材料社区的模型。

This grant provides support for the acquisition of a Beowulf cluster for computational materials research, education and student training at the University of Michigan. Due to phenomenal advances in computer power and accessibility, simulation now guides discovery and understanding of materials processes in nanoscience and technology, microelectronics, polymer processing, photonics, designed materials, autonomic materials systems, and many other technologically important areas. The simulation techniques required to solve materials problems are inherently multidisciplinary, involving length and time scales that span many orders of magnitude. As such, they require the combined use of disparate computational methodologies originally conceived for specific applications in particular disciplines. Expertise in these different techniques tends to remain isolated within their fields of origin. These disciplinary boundaries must be eliminated in order to develop scale-spanning simulation methodologies that coordinate and integrate complementary computational methods and thereby facilitate the investigation of problems too complex to be tackled by a single technique. The goal is three-fold: (1) to foster interdisciplinary and innovative research and education in computational materials at the University of Michigan; (2) to lead in the development and application of cutting-edge, scale-spanning simulation methods for materials research; and (3) to educate and train students in computational methods that span the length and time scales from molecules to macroscopic properties for materials discovery, design, and development. We anticipate our graduates to be vigorously recruited by industry, government laboratories, and universities. A center format is imperative for this endeavor because of the need to pool expertise in diverse computational methodologies. Through the various programs planned, this center will provide a forum for interaction among faculty and students to promote the cross-fertilization between the many areas of specialization that drive advances in computational materials research. This cluster will permit the simultaneous use by multiple users for education, training, and research, and will allow simulations of materials and materials processes over an order of magnitude larger in some instances than previously attainable. Such improvements in scale are crucial to advancing fundamental understanding and scientific discovery. A broader impact of this project is achieved both through the development of innovative computational methodologies and the training of highly skilled computational scientists and engineers, capable of addressing the growing need for the advancement of materials technologies in today's society. Our vision is to be a highly visible leader in research and education in computational materials science and engineering, developing simulation techniques, curricula, and outreach programs at Michigan that will serve as models for the materials community around the world.