A Performance Comparison of Hierarchical Ring- and Mesh-Connected Multiprocessor Networks

This paper compares the performance of hierar-chical ring- and mesh-connected wormhole routed shared memory multiprocessor networks in a simu-lation study. Hierarchical rings are interesting al-ternatives to meshes since i) they can be clocked at faster rates, ii) they can have wider data paths and hence shorter message sizes, iii) they allow addition and removal of processing nodes at arbitrary locations, iv) their topology allows natural exploitation in the spatial locality of application memory access patterns, and v) their topology allows ecient implementation of broadcasts. Our study shows that for workloads with little lo-cality, meshes scale better than ring networks because ring-based systems have limited bisection bandwidth. However, for workloads with some memory access lo-cality, hierarchical rings outperform meshes by 20-40% for system sizes of up to 128 processors. Even with poor access locality, hierarchical rings will outperform meshes for these system sizes if the mesh router buers are only 1- it large, and they will outperform meshes in systems with less than 36 processors regardless of mesh router buer size.