Related Research

Unstructured mesh generation strategies can be characterized as being essentially scalar and having a large variation in the number of operations required during each step of the calculation. The key to a successful parallelization of these strategies is based on a domain partitioning with an independent discretization of the individual subdomains. An efficient parallel algorithm requires a balance of the work between the processors while maintaining the interprocessor communication at a minimum. However, the domain decomposition for mesh generation algorithms is not straightforward because the only knowledge available at the beginning of the discretization process is the geometric model, the complexity of which has little or no relationship to the work required to generate the mesh. The parallelization of mesh generation techniques is therefore a non-trivial task and only few approaches for the parallel mesh generation have been developed so far. These approaches are frequently ``only'' extensions of their sequential versions into the parallel environment subjected to some modifications taking into account the domain decomposition and load balancing. Moreover, most of these approaches are exclusively applying to the generation of triangular and tetrahedral meshes (unless the conversion to quadrilateral or hexahedral elements is employed) because these techniques are better established in the sequential environment.

*Daniel Rypl
2005-12-07*