Parallel Computing on Distributed Memory Multiprocessors

Advances in microelectronic technology have made massively
parallel            computing a reality and triggered an outburst of
research activity in         parallel processing architectures and
algorithms. Distributed memory            multiprocessors - parallel
computers that consist of microprocessors            connected in a
regular topology - are increasingly being used to                  solve
large problems in many application areas. In order to use
these   computers for a specific application, existing
algorithms need to be            restructured for the architecture and
new algorithms developed. The            performance of a computation
on a distributed memory multiprocessor is         affected by the
node and communication architecture, the                              interconnection
network topology, the I/O subsystem, and the                        parallel
algorithm and communication protocols. Each of these
parametersis a complex problem, and solutions require an
understanding of the            interactions among them.
This book is based on the papers presented at the   NATO
Advanced Study Institute held at Bilkent University, Turkey,
in      July 1991. The book is organized in five parts: Parallel
computing               structures and communication, Parallel numerical
algorithms, Parallel         programming, Fault tolerance, and
Applications and algorithms.