Parallel Computing on Distributed Memory Multiprocessors

Advances in microelectronic technology have made massivelyparallel computing a reality and triggered an outburst ofresearch activity in parallel processing architectures andalgorithms. Distributed memory multiprocessors - parallelcomputers that consist of microprocessors connected in aregular topology - are increasingly being used to solvelarge problems in many application areas. In order to usethese computers for a specific application, existingalgorithms need to be restructured for the architecture andnew algorithms developed. The performance of a computationon a distributed memory multiprocessor is affected by thenode and communication architecture, the interconnectionnetwork topology, the I/O subsystem, and the parallelalgorithm and communication protocols. Each of theseparametersis a complex problem, and solutions require anunderstanding of the interactions among them. This book is based on the papers presented at the NATOAdvanced Study Institute held at Bilkent University, Turkey, in July 1991.The book is organized in five parts: Parallelcomputing structures and communication, Parallel numericalalgorithms, Parallel programming, Fault tolerance, andApplications and algorithms.