Processor Netorks and Aspects of the Mapping Problem [Paperback]

This 1991 book discusses how to use processors simultaneously in order to solve a single problem rather than any specific application.This 1991 book discusses the general case of how to use processors simultaneously in order to solve a single problem rather than any specific application, and develops a theory independent of particular architectures. It is intended to be a fundamental treatment of the relevant subjects and is aimed at computer scientists and graduate students.This 1991 book discusses the general case of how to use processors simultaneously in order to solve a single problem rather than any specific application, and develops a theory independent of particular architectures. It is intended to be a fundamental treatment of the relevant subjects and is aimed at computer scientists and graduate students.Processor networks consisting of a thousand and more processing nodes that interact by sending messages over a communications network are feasible today. They can only meet the market's demand for high-speed, low-cost computing when adequate facilities for the implementation of distributed programs--programs consisting of a collection of processes that communicate and synchronize via the exchange of messages--are offered. This book investigates some of these facilities by viewing both the processor network and the distributed program as a graph. Dr. Hilbers begins by introducing distributed computing with graph theory, and considers processor networks and their price/performance ratios. He goes on to look at obtaining homogeneous distributions of work over networks and considers examples. The author closes by discussing message routing within a processor network.Preface; 1. Introduction; 2. Processor networks; 3. A new operation on graphs; 4. Homogenous mappings; 5. Mappings of binary tree computation graphs; 6. Mappings on torus-connected graphs; 7. Deadlock-free message routing in processor networks; References; Index.