Parallel and distributed computation: numerical methods
Parallel and distributed computation: numerical methods
Data communication in hypercubes
Journal of Parallel and Distributed Computing
Multicast in hypercube multiprocessors
Journal of Parallel and Distributed Computing
An Adaptive and Fault Tolerant Wormhole Routing Strategy for k-ary n-cubes
IEEE Transactions on Computers
Deadlock-free multicast wormhole routing in multicomputer networks
ISCA '91 Proceedings of the 18th annual international symposium on Computer architecture
Unicast-Based Multicast Communication in Wormhole-Routed Networks
IEEE Transactions on Parallel and Distributed Systems
Hi-index | 0.24 |
In this paper we present a systematic method of configuring group-multicast rings with randomly selected nodes on binary n-cube multiprocessors. The method configures rings only with disjoint links so that the selected nodes can perform multicast communication without messages collision. Binary n-cube multiprocessors are typically one-port architectures, which allow a processor node to send or receive only one message at a time. Being capable of pipelining messages throughout nodes, rings offer a practical solution to perform group multicast operations on one-port architectures. The basic idea behind our proposed method is to repeatedly merge small rings to form larger rings. Once a multicast ring is configured, messages can circulate around the ring without incurring link contentions. We begin with introducing a basic network model which specifies path routing on n-cubes. Then we present some preliminary results regarding properties of disjoint paths. The concept of the proposed method is described. Finally, the method is formally presented and its correctness is verified. The proposed method presents two unique features. First, it guarantees that the physical distance between adjacent nodes is limited to n links on an n-cube. Secondly, it uses only a single dimension order to route messages across the links towards their respective destinations.