Parallel implementation of an adaptive scheme for 3 D unstructuredgrids

Dynamic mesh adaption on unstructured grids is a powerful tool for computing unsteady ows that require local grid modiications in order to eeciently resolve solution features. For such ows, the coarsening/reenement step must be completed every few time steps, so its eeciency must be comparable to that of the ow solver. For this work, we consider an edge-based adaption scheme that has shown good single processor performance on a Cray Y-MP and C90. We report on our experience porting this code to an SGI Power Challenge system and parallelizing it for shared-memory symmetric multiprocessor architectures.