Parallel Computation: Models And Methods

Transformational programming and parallel computation are two emerging fields that may ultimately depend on each other for success. Perhaps because ad hoc programming on sequential machines is so straightforward, sequential programming methodology has had little impact outside the academic community, and transformational methodology has had little impact at all. However, because ad hoc programming for parallel machines is so hard, and because progress in software construction has lagged behind architectural advances for such machines, there is a much greater need to develop parallel programming and transformational methodologies. Parallel Algorithm Derivation and Program Transformation stimulates the investigation of formal ways to overcome problems of parallel computation, with respect to both software development and algorithm design. It represents perspectives from two different communities: transformational programming and parallel algorithm design, to discuss programming, transformational, and compiler methodologies for parallel architectures, and algorithmic paradigms, techniques, and tools for parallel machine models. Parallel Algorithm Derivation and Program Transformation is an excellent reference for graduate students and researchers in parallel programming and transformational methodology. Each chapter contains a few initial sections in the style of a first-year, graduate textbook with many illustrative examples. The book may also be used as the text for a graduate seminar course or as a reference book for courses in software engineering, parallel programming or formal methods in program development.

Computing has moved away from a focus on performance-centric serial computation, instead towards energy-efficient parallel computation. This provides continued performance increases without increasing clock frequencies, and overcomes the thermal and power limitations of the dark-silicon era. As the number of parallel cores increases, we transition into the many-core computing era.

Computing has moved away from a focus on performance-centric serial computation, instead towards energy-efficient parallel computation. This provides continued performance increases without increasing clock frequencies, and overcomes the thermal and power limitations of the dark-silicon era. As the number of parallel cores increases, we transition into the many-core computing era.

This book is a compilation of a selected subset of research articles presented at the Eighth INFORMS Computing Society Conference, held in Chandler, Arizona, from January 8 to 10, 2003. The articles in this book represent the diversity and depth of the interface between ORiMS (operations research and the management sciences) and CS/AI (computer science and artificial intelligence ). This volume starts with two papers that represent the reflective and integrative thinking that is critical to any scientific discipline.

The first seven chapters use R for probability simulation and computation, including random number generation, numerical and Monte Carlo integration, and finding limiting distributions of Markov Chains with both discrete and continuous states. Applications include coverage probabilities of binomial confidence intervals, estimation of disease prevalence from screening tests, parallel redundancy for improved reliability of systems, and various kinds of genetic modeling.