Evaluating the impact of locality on the performance of large-scale SCI multiprocessors

Hierarchical ring-based multiprocessor systems are attractive and enjoy sev eral advantages over other type of systems. They ensure unique paths betwee n nodes, simple node interfaces and simple cross-ring connections. Furtherm ore, employing point-to-point links allows the system to run at high clock rate which increases bandwidth and decreases latency. This paper investigat es the performance of hierarchical ring-based shared-memory multiprocessors . Rings in the hierarchy are composed of point-to-point, unidirectional lin ks and apply the Scalable Coherent Interface (SCI) protocol. We pay special emphasis on the impact of locality on processor and interconnection design issues such as number of outstanding requests, and ring topology. We find that in order to exploit the power of hierarchical multiprocessors an accur ate and appropriate model of locality must be used. Hierarchical multiproce ssors that are well balanced (uniform) tend to provide lower latency and hi gher system throughput. For non-uniform systems, high degree of locality is required for the hierarchies to perform well. However, restricting the num ber of outstanding transactions per processor is important in decreasing pa ckets latency and avoiding network contention. (C) 2001 Elsevier Science B. V. All rights reserved.