High Performance Communications In Processor Networks

24 August 2005

conference paper
Published by Institute of Electrical and Electronics Engineers (IEEE)

No. 10636897,p. 150-157
https://doi.org/10.1109/isca.1989.714549

Abstract

In order to provide an arbitrary and fully dynamic connectivity in a network of processors, transport mechanisms must be implemented, which provide the propagation of data from processor to processor, based on addresses contained within a packet of data. Such data transport mechanisms must satisfy a number of requirements - deadlock and livelock freedom, good hot-spot performance, high throughput and low latency. This paper proposes a solution to these problems, which allows deadlock free, adaptive, high throughput packet routing to be implemented on networks of processors. Examples are given which illustrate the technique for 2-D array and toroidal networks. An implementation of this scheme on arrays of transputers is described. The scheme also serves as a basis for a very low latency routing strategy named the mad postman, a detailed implementation of which is described here as well.

Keywords

This publication has 6 references indexed in Scilit:

Programming with active data
Published by Springer Nature ,1989
Design of SIMD microprocessor array
IEE Proceedings E Computers and Digital Techniques, 1989
Deadlock-Free Message Routing in Multiprocessor Interconnection Networks
IEEE Transactions on Computers, 1987
The cosmic cube
Communications of the ACM, 1985
A DAG-Based Algorithm for Prevention of Store-and-Forward Deadlock in Packet Networks
IEEE Transactions on Computers, 1981
Virtual cut-through: A new computer communication switching technique
Computer Networks (1976), 1979