SCREMS: High Performance Computing and Visualization

PI建议购买配备GPU卡的多节点集群系统，以大大改善佛罗里达州立大学数学系的高性能计算环境。该集群将致力于高性能计算（HPC）和可视化的研究，并为教师和学生提供HPC培训。拟议的研究项目将利用集群系统的力量，范围包括液晶聚合物和纳米复合材料的多尺度建模和计算、生物材料和流体模拟、流体流动和多相现象的计算研究以及使用GPU的可视化。非线性波在各种介质中的传播。 这些项目中的每一个都针对特定的材料或物理系统、一种技术或一类数学方程解决了重要问题，并旨在利用计算最大限度地解决这些问题。此次收购将使我们能够利用HPC的尖端技术创造新的研究能力，从而对应用和计算数学的研究和教育产生影响。佛罗里达州立大学的PI一直在进行理解和开发高性能材料，流体流动模拟，多相流运动，生物材料建模和仿真以及各种介质中的非线性波传播，例如光纤和液晶材料。这些前沿的研究活动需要前沿的高性能计算技术的支持。最近，计算社区正在投入大量资金，将最初为游戏目的设计的图形卡变成计算机系统上除了CPU之外的并行数据处理单元。随着高性能计算技术的发展，利用图形卡的数据处理能力，将使计算机用户在传统的CPU数据处理能力之外，拥有更强大的处理大量数据的能力。建议的研究项目，由建议的集群系统配备的图形卡计算能力，将使人们能够开发高性能材料，了解生物材料的功能，如何非线性光学可以应用于产生功率波，以及如何水下爆炸将影响海军船舶等。 它将大大提高我们对所研究问题的基础科学的理解，并对未来的计算研究和计算科学家的培养产生影响。

The PIs propose to acquire a multi-node cluster system equipped with GPU cards to vastly improve the high performance computing environment in the Department of Mathematics at the Florida State University. The cluster will be dedicated to research in high performance computing (HPC) and visualization as well as providing training on HPC for faculty and students. The proposed research projects that will take advantage of the power of the cluster system range from multiscale modeling and computation of liquid crystalline polymers and nanocomposites materials, biomaterials and fluid simulation, computational study of fluid flows and multiphase phenomena, and visualization using GPUs to nonlinear wave propagations in various media. Each of these projects addresses important issues for a particular material or physical system, a type of technology, or a class of mathematical equations, and aims to resolve the issues to the fullest extent using computation. The acquisition will enable us to create new capabilities in research using the cutting-edge technology in HPC and thereby impact on research as well as education in applied and computational mathematics.The PIs at Florida State University have been conducting research in understanding and developing high performance materials, fluid flow simulation, multiphase flow motions, biomaterials modeling and simulation as well as nonlinear wave propagation in various media such as fiber optics and liquid crystal materials. These cutting edge research activities demand the support of the cutting-edge high performance computing technologies. Recently, the computing community is investing heavily in turning the graphic card originally designed for gaming purposes into a parallel data processing unit in addition to the CPU on computer systems. The developing high performance computing technology using graphic card data processing capability will provide computer users much improved computing power to process large amount of data in addition to the traditional CPU data processing. The proposed research projects aided by the proposed cluster system equipped with the graphic card computing capability will enable ones to develop high performance materials, to understand how biomaterials function, how nonlinear optics can be applied to produce power waves, and how underwater explosion will impact on naval ships etc.. It will significantly improve our understanding on the fundamental science of the problems to be studied and impact on the future research in computation and training of computational scientists.