PARALLEL ADAPTIVE MESH GENERATION AND DECOMPOSITION

An important class of methodologies for the parallel processing of com putational models defined on some discrete geometric data structures ( i.e. meshes, grids) is the so called geometry decomposition or splitti ng approach. Compared to the sequential processing of such models, the geometry splitting parallel methodology requires an additional comput ational phase. It consists of the decomposition of the associated geom etric data structure into a number of balanced subdomains that satisfy a number of conditions that ensure the load balancing and minimum com munication requirement of the underlying computations on a parallel ha rdware platform. It is well known that the implementation of the mesh decomposition phase requires the solution of a computationally intensi ve problem. For this reason several fast heuristics have been proposed . In this paper we explore a decomposition approach which is part of a parallel adaptive finite element mesh procedure. The proposed integra ted approach consists of five steps. It starts with a coarse backgroun d mesh that is optimally decomposed by applying well known heuristics. Then, the initial mesh is refined in each subdomain after linking the new boundaries introduced by its decomposition. Finally, the decompos ition of the new refined mesh is improved so that it satisfies the obj ectives and conditions of the mesh decomposition problem. Extensive ex perimentation indicates the effectiveness and efficiency of the propos ed parallel mesh and decomposition approach.