Parallel Prefix Computation and Sorting on a Recursive Dual-Net
The purpose of the Interconnection Networks (INs) is to connect processor/memory boards together to form a parallel or distributed system. In massively parallel computer systems, the interconnection networks play a crucial role in issues such as communication performance, hardware cost, computational complexity, and fault-tolerance. Much research has been reported in the literature on interconnection networks, which can be used to construct parallel computers of large scale. The following two categories have attracted great research attention. One is the hyper-cube like family that has the advantage of short diameters for high-performance computing and efficient communication. The other is the family of 2D/3D meshes or tori that has the advantage of small and fixed node-degrees and easy implementation.