Abstract: Standard ML is a functional programming language used extensively in universities and increasingly in industry. This paper discusses a concurrency mechanism which has been implemented in the Poly/ML implementation of Standard ML and has been used on uniprocessors and shared memory multiprocessors. It is now being implemented on a distributed network of Unix workstations. Each of these implementations is described.
The aim of this work is to produce a distributed system that will allow large ML programs to be run on a network of processors. Although eventually such a network might be a closely-coupled network of transputers, the initial design is intended for the sort of system that many organisations have, namely personal workstations on a local network. Making use of these out of office hours will provide a substantial improvement in the computing power available.Previous | Index | Next