Task Scheduling for Parallel Systems

A new model for task scheduling that dramatically improves theefficiency of parallel systems Task scheduling for parallel systems can become a quagmire ofheuristics, models, and methods that have been developed over thepast decades. The author of this innovative text cuts through theconfusion and complexity by presenting a consistent andcomprehensive theoretical framework along with realistic parallelsystem models. These new models, based on an investigation of theconcepts and principles underlying task scheduling, take intoaccount heterogeneity, contention for communication resources, andthe involvement of the processor in communications. For readers who may be new to task scheduling, the firstchapters are essential. They serve as an excellent introduction toprogramming parallel systems, and they place task scheduling withinthe context of the program parallelization process. The author thenreviews the basics of graph theory, discussing the major graphmodels used to represent parallel programs. Next, the authorintroduces his task scheduling framework. He carefully explains thetheoretical background of this framework and provides severalexamples to enable readers to fully understand how it greatlysimplifies and, at the same time, enhances the ability toschedule. The second half of the text examines both basic and advancedscheduling techniques, offering readers a thorough understanding ofthe principles underlying scheduling algorithms. The final twochapters address communication contention in scheduling andprocessor involvement in communications. Each chapter features exercises that help readers put their newskills into practice. An extensive bibliography leads to additionalinformation for further research. Finally, the use of figures andexamples helps readers better visualize and understand complexconcepts and processes. Researchers and students in distributed and parallel computersystems will find that this text dramatically improves theirability to schedule tasks accurately and efficiently.