CAREER: Integrated Research & Education in the Use of HPC for Realistic Simulations

职业：综合研究

• 批准号: 9702947
• 负责人: Abani Patra
• 金额: $ 23.26万
• 依托单位: SUNY at Buffalo
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1997
• 资助国家: 美国
• 起止时间: 1997-03-15 至 2002-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9702947&HistoricalAwards=false
• 关键词: CAREER; Integrated; Research; & Education

Research in this project seeks to address issues arising in two computationally challenging problems with great economic potential, namely the analysis of dental implants and supporting bone structure, and the modeling of complex time-dependent 3D viscous flows. These applications will drive development of an integrated problem solving environment that can subsequently be used to solve many similar problems. The overall goal of the research is the simultaneous optimization of the computational process with adaptive hp discretizations and the use of the most powerful HPC hardware available, namely massively parallel distributed memory parallel architectures. This combination of technologies should allow realistic computer simulations of the complex time dependent behavior inherent in both of these problems, for the first time. However, this goal needs several advances in the enabling technologies required for parallel adaptive methods - data structures, mesh partitioning, inter-processor communications, error estimation, adaptive mesh design strategies, and fast and stable solvers for adaptive grids in multiprocessor environments.

该项目的研究旨在解决两个具有巨大经济潜力的具有计算挑战性的问题，即牙种植体和支撑骨结构的分析，以及复杂的随时间变化的三维粘性流动的建模。这些应用程序将推动集成问题解决环境的开发，该环境随后可用于解决许多类似的问题。该研究的总体目标是通过自适应高性能离散化和使用最强大的高性能计算硬件（即大规模并行分布式内存并行架构）同时优化计算过程。这两种技术的结合将首次允许对这两个问题中固有的复杂时间依赖行为进行真实的计算机模拟。然而，这一目标需要在并行自适应方法所需的支持技术方面取得若干进展——数据结构、网格划分、处理器间通信、误差估计、自适应网格设计策略以及多处理器环境下自适应网格的快速稳定求解器。