Ada For Multi Microprocessors

Ada for Multi-Microprocessors

Multi-processor systems are becoming more prevalent due to the many advantages which they offer over single-processor systems. A multi-microprocessor system can smoothly accommodate expansion and increasing levels of integration of separate controllers, facilitating modernization of industrial plants without large capital outlay. This book deals with the ways in which Ada can be used for such distributed systems. The main part of the book is devoted to the issue of how to construct and run an Ada program for a variable target configuration of several microcomputers, interconnected through shared memories, multi-access busses, local area networks or end-to-end lines. It is recognized that Ada and current Ada Programming Support Environments (APSEs) do not address distributed targets, and different approaches are therefore considered for coping with distribution without changing the language or unduly restricting the way in which it is used. The approach which emerges as the most natural is one where the designer is required to cluster tightly coupled Ada tasks into 'virtual nodes'. The implications of this approach for the user and the APSE are examined in detail, and further implications concerning use of the multi-microprocessor approach to achieve reliability and extensibility are also studied.

Multi-microprocessors

A Review of Ada Tasking

Ada* is unquestionably one of the most significant programming languages to emerge in the last decade. The manner of its inception and support by the US Department of Defense (DoD) ensures that it will be used extensively for the indefinite future in programming large and complex systems. The growing availability of compilers means that many organisations are already committed to using the language for sizable and significant applications. As a perhaps inevitable result of its design goals, Ada is a "large" language. It has Pascal-like control and type constructs; a mechanism for exception handling; a package structure for information hiding, decomposition and separate compilation; facilities for low-level programming; and a tasking model of concurrency. It is perhaps this last area that has generated most debate, criticism and disagreement. The purpose of this book is to review the tasking model in the light of the extensive analysis and comment which has appeared in the literature. The review is necessarily wide-ranging, including discussion of - Ada as a general purpose concurrent programming language, - Ada as a language for embedded and distributed systems, - implementation issues, with particular reference to distributed systems, - formal semantics, specification and verification, - proposed language modifications. By consolidating this discussion within the confines of a single review, potential users of the tasking facility are enabled to familiarise themselves with all the factors which may impinge upon the performance, reliability and correctness of their software. The book also provides a focus for any debate on modifications to the Ada language, or developments from it.