Abstracts for the Conference on
Grid Adaptivity in Computational PDEs,
Edinburgh, July 96		Next
Previous
Index

Load-balancing for parallel adaptive unstructured grids

C Walshaw , M Cross & M G Everett

Centre for Numerical Modelling and Process Analysis,
University of Greenwich, London, SE18 6PF, UK
e-mail:

Abstract

The use of adaptive grid techniques on parallel machines can be one of the most efficient ways to solve large PDE problems. In particular, if the computational mesh is unstructured or irregular, completely general geometries and complex behaviour can be readily modelled and, in principle, the inherent sparsity of many such problems can be exploited to obtain excellent parallel efficiencies. However such codes can give rise to a series of meshes in which the position and density of the data points varies dramatically over the course of an integration and which may need to be frequently repartitioned in order that the computational load is evenly balanced and the amount of interprocessor communication is minimised.

In this talk we shall describe a parallel optimisation strategy for solving this dynamic load-balancing problem. The optimisation technique which marries a Kernighan Lin type algorithm with a distributed load-balancing algorithm will be outlined and we will briefly discuss the choice of suitable gain functions. The algorithm works both in serial and in parallel and results indicative of both partition quality and the parallel performance of the partitioning will be presented.

ABSTRACTS
NextPreviousIndex
Last modified Fri Jun 21 19:19:17 GB-Eire 1996 (DBD)