Aligning block permutation methods for topology transformation on computational grids

This Traditional parallel matrix computing algorithms usually perform on the regular resource topologies, such as a mesh. However, the grid resource topology is often irregular. To perform traditional parallel algorithms on a modern grid environment, the transformation method of grid resource topology for a virtual mesh is proposed in this paper. The basic idea of our transformation is based on the aligning block permutation in a virtual mesh. To utilize efficiently the computational and communicational abilities of grid resources, a heuristic search method is used in this transformation. The experiment results show that the aligning block permutation can reduce the time complexity of search tree and the heuristic search can find the block permutation with better utilization computational and communicational abilities of resources in computational grids.