doi:10.1016/S0045-7906(03)00004-1
Copyright © 2003 Elsevier Ltd. All rights reserved.
Compressionless wormhole routing: an analysis for hypercube with virtual channels
References and further reading may be available for this article. To view references and further reading you must
purchase this article.
A. Khonsari and M. Ould-Khaoua
, 
Department of Computing Science, University of Glasgow, Glasgow G12 8RZ, UK
Received 31 May 2001;
revised 29 August 2001;
accepted 5 December 2001. ;
Available online 20 February 2003.
Abstract
Several recent studies have shown that adaptive routing algorithms based on deadlock recovery have superior performance characteristics than those based on deadlock avoidance. Most of these studies, however, have relied on software simulation due to the lack of analytical modelling tools. In an effort towards filling this gap, this paper presents a new analytical model of compressionless routing in wormhole-routed hypercubes. This routing algorithm exploits the tight coupling between wormhole routers for flow control to detect and recover from potential deadlock situations. The advantages of compressionless routing include deadlock-free adaptive routing with no extra virtual channels, simple router design, and order-preserving message transmission. The proposed analytical model computes message latency by determining the message transmission time, blocking delay at each router, multiplexing delay at each network channel, and waiting time in the source before entering the network. The validity of the model is demonstrated by comparing analytical results with those obtained through simulation experiments.
Author Keywords: Multicomputers; Interconnection networks; Fully adaptive routing; Compressionless routing; Virtual channels; Performance modelling
Fig. 1. Examples of n-dimensional hypercube: (a) two-dimensional hypercube, (b) three-dimensional hypercube, and (c) six-dimensional hypercube.
Fig. 2. The node structure in a hypercube.
Fig. 3. The state transition diagram of an
i-hop message in the network.
Fig. 4. Morkov model for virtual channel occupancy probabilities.
Fig. 5. The mean message latency predicated by the analytical model and simulation against the traffic generation rate in the 2-ary 6-cube. The number of virtual channels
V=1, 2, 3, 4, mean message length
M=32 and 64 flits, timeout period equals the mean message length (τ=
M) and mean re-transmission time gap equal the mean message length .
Fig. 6. The mean message latency predicated by the analytical model and simulation against the traffic generation rate in the 2-ary 8-cube. The number of virtual channels
V=1, 2, 3, 4, mean message length
M=32 and 64 flits, timeout period equals the mean message length (τ=
M) and mean re-transmission time gap equal the mean message length .
Corresponding author. Tel.: +44-141-3306056; fax: +44-141-3304913