Introduction

Abstract

Parallel processing is a technique to design high performance computers, which are required to solve “grand challenge” problems. There are also several other reasons for the increased interest in parallel processing. In the first section, we look at the reasons for this interest. Unlike uniprocessor systems, parallel systems can be designed using a variety of architectures. A brief overview of parallel architectures is given in Section 1.2. Managing resources efficiently is important in parallel systems in order to benefit from parallel processing. In this book, we are concerned with one aspect of resource management: job scheduling. We introduce job scheduling in Section 1.3. Starting with Chapter 3, we provide a detailed discussion of various job scheduling policies. Performance of parallel job scheduling policies depends on the type of software architecture employed. We briefly describe some of the basic software architectures in Section 1.4. We conclude the chapter with an overview of the book.