Repartitioning Unstructured Adaptive Meshes

We present a new parallel repartitioning algorithm for adaptive finite-element meshes that significantly reduces the amount of data that needs to move between processors in order to rebalance a workload after mesh adaptation (refinement or coarsening). These results derive their importance from the fact that the time to migrate data can be a large fraction of the total time for the parallel adaptive solution of partial differential equations.

• Type: Conference
• Year: 2000
• Where: IPPS
• Authors: José G. Castaños, John E. Savage
• Keywords: Adaptive finite-element Meshes, Distributed And Parallel Computing, IPPS 2000, Parallel Adaptive Solution, Parallel Repartitioning Algorithm