Dynamic load balancing with an improved with an improved spectral bisection algorithm

The efficient parallel execution of grid-oriented scientific calculations requires the partitioning of the grid in such a way that the work load is equally distributed over the processors of the parallel machine and that communication among the processors is minimized. For unstructured static grids, good partitions are obtained by applying the recursive spectral bisection heuristic to the interdependency graph of the grid. We describe how even in case of dynamically changing grids, grid-oriented problems can be formulated as graph partitioning problems for the purpose of load balancing. We further describe an alternative spectral bisection algorithm that yields better partitions than the standard algorithm for the graphs that model dynamic load balancing problems.<
>