MPI (Message Passing Interface) is a specification for a standard library for message passing that was defined by the MPI Forum, a broadly based group of parallel computer vendors, library writers, and applications specialists. Multiple implementations of MPI have been developed. In this paper, we describe MPICH, unique among existing implementations in its design goal of combining portability with high performance. We document its portability and performance and describe the architecture by which these features are simultaneously achieved. We also discuss the set of tools that accompany the free distribution of MPICH, which constitute the beginnings of a portable parallel programming environment. A project of this scope inevitably imparts lessons about parallel computing, the specification being followed, the current hardware and software environment for parallel computing, and project management; we describe those we have learned. Finally, we discuss future developments for MPICH, including those necessary to accommodate extensions to the MPI Standard now being contemplated by the MPI Forum. 1

For almost a decade we have been working at developing and using template-based models for coarse-grained parallel computing. Our initial system, FrameWorks, was positively received but had a number of shortcomings. The Enterprise parallel programming environment evolved out of this work, and now, after several years of experience with the system, its shortcomings are becoming evident. This paper outlines our experiences in developing and using the two parallel programming systems. Many of our observations are relevant to other parallel programming systems, even though they may be based on different assumptions. Although template-base models have the potential for simplifying the complexities of parallel programming, they have yet to realize these expectations for high-performance applications. 1 Introduction Along with the growing interest in parallel and distributed computing, there has been a corresponding increase in the development of models, tools and systems for parallel progra...

This report addresses the issues arising from the use of parallel machines and considers the various techniques used by members of the consortium in this context. Before considering the algorithms in detail, we describe, in section 2, the main types of parallel architecture and survey various attempts at providing a taxonomy. Then, in section 3, we address the difficult issue of the measurement of processor performance in order to quantify any enhancement obtained by implementing an algorithm in parallel. Section 4 presents the main features of in general, and PVM ( ) in particular. The latter is an application that is used to generate distributed versions of sequential algorithms for use on networks of workstations. The parallel implementations of the GA toolkit, , and the associated simulated annealing toolkit, , both developed at UEA, have been produced using PVM. Exact algorithms will always find the optimal solution to a problem given enough time and space. Subject to these constraints, they must always be the preferred method of solution. In practice, the time and space constraints can prevent the use of an exact algorithm and thus the potential of parallelism to reduce these factors becomes an important factor. Total enumeration is embarassingly parallel. With processors it is reasonable to expect an-fold reduction in time to undertake such a thorough search. Such a saving is seldom sufficient to make the method viable so we will concentrate on other exact methods here. In section 5, we review parallel branchand -bound, reprinting a survey paper written by the UEA partners in the consortium and previously published in [1]. Because of the interest in interior point methods for the CALMA project and its widely cited potential for parallelisation, this provides the ...