Next: Tuple-space Kernel.

Distributed Tuple-Spaces for Parallel Image Processing

A.M. Wood
University of York

14 April 1996

This study was focussed on the use and implementation of the tuple-space model of concurrency, which originated at Yale with the LINDA system - a co-ordination language for expressing parallel processing, whose strength lies in its ability to describe parallel algorithms without reference to any specific computer architecture. The key feature of LINDA is that it provides inter-process co-ordination via virtual shared memories (tuple-spaces) which are accessed associatively. By incorporating the LINDA primitives and tuple-spaces (`bags' of tuples) into a programming language, algorithms developed using the model can be more easily ported between different parallel computers. It is this ability to abstract away from particular computational structures that makes the tuple-space model a particularly attractive vehicle for parallel algorithm and architecture research.

The key outcomes of this project were:


wood@minster.york.ac.uk
Mon May 6 13:40:14 BST 1996