The Data Diffusion Machine is a scalable virtual shared memory architecture. A hierarchical network is used to ensure that all data can be located in a time bounded by O(logp), where p is the number of processors. The DDM hierarchy requires a high degree of connectivity between clusters of nodes, which can be provided with point-topoint links. For large machines however the wiring will be complex. In this article we discuss the implementation of such networks, and develop three alternative implementations. The base level performance of each alternative has been measured on an emulator of the DDM. The final solution collapses the physical hierarchy, and we show that this does not affect the performance, while clearly simplifying the design. It demonstrates that with the use of crossbar routers we can make a cheap, scalable and high performance implementation of the DDM.
Henk L. Muller, Paul W. A. Stallard, David H. D. W