Analysis of modern aerospace vehicles requires the computation of flowfields about complex 3-D geometries composed of regions with varying spatial resolution requirements. Overset grid methods allow the use of proven structured grid flow solvers to address the twin issues of geometrical complexity and the resolution variation by decomposing the complex physical domain into a collection of overlapping subdomalns. This flexibility is accompanied by the need for irregular intergrid boundary communication among the overlapping component grids. This study investigates a strategy for implementing such a static overset grid implicit flow solver on distributed memory, M1MD computers; i.e., the 128 node Intel iPSC/860 and the 208 node Intel Paragon. Performance data for two composite grid configurations characteristic of those encountered in present day aerodynamic analysis are also presented. Introduction Justifying Parallelization: The complexity of Computational Fluid Dynamics (CFD) simulations attempted at present is very closely related to the sustained CPU performance of the readily available computer resources. Simplified, 2-D flow analysis simulations can be carried out using conventional high performance workstations on a regular basis. However, 3-D unsteady, viscous flow analysis still requires the very best of computing hardware. Most of the current generation of vector supercomputers such as the Cray C-90 and the NEC SX-3 are fully capable of providing the compute resources required for such simulations. However the high cost of such machines and their consequent limited availability have spurred efforts aimed at seeking more cost effective approaches to high performance, numericallyintensive computing. The most prominent among a number of such computational strategies being currently investigated underthe umbrellaof the national High Performance Computing and Communications Program (HPCCP) is the one based on the exploitation of the relatively high degree of concurrency and the spatial data locality inherent in numerical algorithms used for aerodynamic simulations. Under these conditions, distributed memory, multiple instruction, multiple data (DM-MIMD) computers offer excellent long-term prospects for greatly increased computational speed and memory at a cost that may render the 3-D flow analysis of complex shapes on a routine basis more affordable. Among the recent advances in computer hardware technologies that lend credence to such expectations are the advent of mass-produced high-performance Reduced Instruction Set Computing (RISC) microprocessor chips, high density Dynamic Random Access Memory (DRAM) chips and high-speed interconnect networks that are easily scalable to the level of hundreds of nodes. The essential remaining ingredient required for the success of this mode of computing is the development and implementation of underlying numerical algorithms in a manner that is conducive to retaining high parallel efficiencies when the number of processors used range at least in the low hundreds. This of...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.