MRI: Acquisition of 100TF Graphics Processor Laboratory for Multiscale/Multiphysics Modeling

MRI：收购 100TF 图形处理器实验室用于多尺度/多物理场建模

• 批准号: 0923018
• 负责人: Mark Robbins
• 金额: $ 38.15万
• 依托单位: Johns Hopkins University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2009
• 资助国家: 美国
• 起止时间: 2009-08-01 至 2013-07-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0923018&HistoricalAwards=false
• 关键词: MRI; Acquisition; 100TF; Graphics; Processor

Complex systems that involve a wide range of length and time scales and different physical processes are at the forefront of science and engineering. New computational hardware and algorithms are needed to address these challenging problems. The goal of this award is to acquire a compute cluster built with graphics processor units that have the potential to dramatically accelerate calculations. The cluster will allow NSF funded research teams to develop new multiscale/multiphysics algorithms and implement them for graphics processors. Applications to solids will include modeling and optimization of the mechanical properties of materials with heterogeneous structures, connecting atomistic dynamics to the macroscopic response of disordered metals and polymers, and fracture under conditions from slow laboratory loads to planetary impact. The facility will also enable studies of fluids that address temporal and spatial complexity in turbulent flows, water waves and air-sea interactions, and multiphase flow of fluids and particles in porous media. Every member of society is affected by decisions or predictions based on computational simulations of critical processes, such as infrastructural integrity, energy production, and environmental processes. This facility will provide a platform for the development and application of new modeling paradigms and techniques for such complex problems. The facility will serve as the primary "laboratory" for the NSF-Integrative Graduate Education and Research Traineeship program on "Modeling Complex Systems - the Scientific Basis of Coupling Multi-Physics Models at Different Scales." It will play an essential role in the training of the students and postdoctoral scholars associated with this program, producing a cadre of computational scientists who can lead the application of new methods for modeling complex systems.

复杂系统涉及广泛的长度和时间尺度以及不同的物理过程，处于科学和工程的前沿。需要新的计算硬件和算法来解决这些具有挑战性的问题。该奖项的目标是获得一种采用图形处理器单元构建的计算集群，该处理器单元具有显著加速计算的潜力。该集群将允许NSF资助的研究团队开发新的多尺度/多物理算法，并将其应用于图形处理器。固体的应用将包括对具有不同结构的材料的机械性能进行建模和优化，将原子动力学与无序金属和聚合物的宏观响应联系起来，以及在从缓慢的实验室载荷到行星撞击的条件下断裂。该设施还将使流体研究能够解决湍流、水波和海气相互作用中的时间和空间复杂性，以及多孔介质中流体和颗粒的多相流动。每个社会成员都受到基于关键过程的计算机模拟的决策或预测的影响，如基础设施完整性、能源生产和环境过程。这一设施将为开发和应用针对此类复杂问题的新的建模范例和技术提供一个平台。该设施将作为美国国家科学基金会“复杂系统建模--在不同尺度上耦合多个物理模型的科学基础”的综合研究生教育和研究培训计划的主要“实验室”。它将在与该项目相关的学生和博士后学者的培训中发挥重要作用，培养一支能够领导应用新方法对复杂系统建模的计算科学家队伍。