U.S.-France (INRIA) Cooperative Research: Domain Decomposition and Multilevel Techniques in Large-Scale Parallel Computing

这个为期三年的奖项支持美国和法国在科罗拉多大学（丹佛和博尔德）、德克萨斯州A&amp；M、INRIA rocqucourt（法国国家计算机科学与应用数学研究所）和ONERA的研究团队在应用数学和先进科学计算方面的合作研究。研究的目的是探讨流体流动和结构动力学的并行数值模拟的区域分解和多层技术。这种方法背后的思想是在并行计算环境中开发新的方法，以允许对更复杂的工程问题进行建模。他们将重点研究四个应用：(1)流动向湍流的转变；(2)分区耦合流动建模；(3)输运现象；(4)流固耦合。所提出的领域出现在许多民用和航空航天工程应用中。美国研究人员将并行计算、多层次和领域分解技术方面的专业知识带入了这次合作。法国团队在区域分解、流体流动和并行计算方面的经验补充了这一点。他们的共同努力可能会导致大规模并行计算算法的改进。* * *

9310529 McCormick This three-year award supports U.S.-France cooperative research in applied mathematics and advanced scientific computing among U.S. and French research teams at the University of Colorado (Denver and Boulder), Texas A&M, and INRIA Rocquencourt (French National Institute for Research in Computer Science and Applied Mathematics) and ONERA. The objective of the research is to investigate domain decomposition and multilevel techniques for parallel numerical simulation of fluid flow and structural dynamics. The idea behind this approach is to develop new methods in a parallel computing environment to allow modeling of more sophisticated engineering problems. They will focus the research on four applications: (1) flow transition to turbulence; (2) zoning-coupled flow modeling; (3) transport phenomena; and (4) fluid-structure interaction. The proposed areas arise in many civil and aerospace engineering applications. The U.S. investigators bring to this collaboration expertise in parallel computing, multilevel and domain decomposition techniques. This is complemented by the French team's experience in domain decomposition, fluid flow and parallel computing. Their joint efforts could lead to improved algorithms for large scale parallel computing. ***