Abstract Machine Models For Parallel And Distributed Computing

Abstract Machine Models have played a profound though frequently unacknowledged role in the development of modern computing systems. They provide a precise definition of vital concepts, allow system complexity to be managed by providing appropriate views of the activity under consideration, enable reasoning about the correctness and quantitative performance of proposed problem solutions, and encourage communication through a common medium of expression. Abstract Models in Parallel and Distributed computing have a particularly important role in the development of contemporary systems, encapsulating and controlling an inherently high degree of complexity. The Parallel and Distributed computing communities have traditionally considered themselves to be separate. However, there is a significant contemporary interest in both of these communities in a common hardware model; a set of workstation-class machines connected by a high-performance network. The traditional Parallel/Distributed distinction therefore appears under threat.

Printbegrænsninger: Der kan printes kapitelvis.

This volume contains the proceedings from the workshops held in conjunction with the IEEE International Parallel and Distributed Processing Symposium, IPDPS 2000, on 1-5 May 2000 in Cancun, Mexico. The workshopsprovidea forum for bringing together researchers,practiti- ers, and designers from various backgrounds to discuss the state of the art in parallelism.Theyfocusondi erentaspectsofparallelism,fromruntimesystems to formal methods, from optics to irregular problems, from biology to networks of personal computers, from embedded systems to programming environments; the following workshops are represented in this volume: { Workshop on Personal Computer Based Networks of Workstations { Workshop on Advances in Parallel and Distributed Computational Models { Workshop on Par. and Dist. Comp. in Image, Video, and Multimedia { Workshop on High-Level Parallel Prog. Models and Supportive Env. { Workshop on High Performance Data Mining { Workshop on Solving Irregularly Structured Problems in Parallel { Workshop on Java for Parallel and Distributed Computing { WorkshoponBiologicallyInspiredSolutionsto ParallelProcessingProblems { Workshop on Parallel and Distributed Real-Time Systems { Workshop on Embedded HPC Systems and Applications { Recon gurable Architectures Workshop { Workshop on Formal Methods for Parallel Programming { Workshop on Optics and Computer Science { Workshop on Run-Time Systems for Parallel Programming { Workshop on Fault-Tolerant Parallel and Distributed Systems All papers published in the workshops proceedings were selected by the p- gram committee on the basis of referee reports. Each paper was reviewed by independent referees who judged the papers for originality, quality, and cons- tency with the themes of the workshops.

This book focuses on advanced techniques used in the design of efficient parallel programs. It presents a wide variety of different models of parallel and distributed computation and applications of these models to the design of efficient algorithms to solve numerical and non-numerical problems. It contains general and specific texts about advanced algorithms for parallel computation and gathers together the state-of-the-art on parallelism with contributions from researchers actively working with parallel computation. Its chapters cover a broad variety of models, ranging from abstract shared-memory models such as PRAM to more realistic models of distributed memory, including CGM, LogP and BSP. To the best of our knowledge, no other book in the literature covers state-of-the-art about parallel models as approached in this special edition. Audience: Practitioners, researchers and graduate students in Computer Science, Mathematics, Engineering and Sciences.

Parallel and distributed computation has been gaining a great lot of attention in the last decades. During this period, the advances attained in computing and communication technologies, and the reduction in the costs of those technolo gies, played a central role in the rapid growth of the interest in the use of parallel and distributed computation in a number of areas of engineering and sciences. Many actual applications have been successfully implemented in various plat forms varying from pure shared-memory to totally distributed models, passing through hybrid approaches such as distributed-shared memory architectures. Parallel and distributed computation differs from dassical sequential compu tation in some of the following major aspects: the number of processing units, independent local dock for each unit, the number of memory units, and the programming model. For representing this diversity, and depending on what level we are looking at the problem, researchers have proposed some models to abstract the main characteristics or parameters (physical components or logical mechanisms) of parallel computers. The problem of establishing a suitable model is to find a reasonable trade-off among simplicity, power of expression and universality. Then, be able to study and analyze more precisely the behavior of parallel applications.

The papers collected in this volume cover a wide range of issues relevant to abstract models, including terminology and concepts for abstract models of computation, models for general purpose parallel computing, declarative models, performance modelling, and special purpose parallel models.The papers originated from the Second Workshop on Abstract Machine Models for Highly Parallel Computers, sponsored by the BCS Parallel Processing Specialist Group. Overall themes of the workshop were the specification, implementation, and application of such models, and the identification of keyissues for future research.

The research described in this book addresses the semantic gap between logic programming languages and the architecture of parallel computers - the problem of how to implement logic programming languages on parallel computers in a way that can most effectively exploit the inherent parallelism of the language and efficiently utilize the parallel architecture of the computer. Following a useful review of other research results, the first project explores the possibilities of implementing logic programs on MIMD, nonshared memory massively parallel computers containing 100 to 1,000 processing elements. The second investigates the possibility of implementing Prolog on a typical SIMD machine, called a Distributed Processor Array. The author's objectives are to define a parallel computational paradigm (the Extended Cellular-Dataflow Model) that can be used to create a Parallel Prolog Abstract Machine as a general starting point for implementing logic programming languages on parallel computers, to exploit the different types of parallelism of these programs, to define an efficient parallel logic machine, to explore the possibilities of implementing logic programming languages on array processors, such as the DAP, to invent parallel implementation techniques for effectively executing Prolog on the DAP, and to define a parallel extension of Prolog that is able to utilize the processor aggregate of the DAP for effectively solving numerical subproblems within logic programs Peter Kacsuk is with Multilogic Computing Ltd. of Budapest, Hungary.