In this paper we describe the difficulties inherent in making accurate, reproducible measurements of message-passing performance. We describe some of the mistakes often made in attempting such measurements and the consequences of such mistakes. We describe mpptest, a suite of performance measurement programs developed at Argonne National Laboratory, that attempts to avoid such mistakes and obta...

We describe a number of early e orts to make use of the Message Passing Interface (MPI) standard in applications, based on an informal survey conducted in May-June, 1994. Rather than a de nitive statement of all MPI development work, this paper addresses initial successes, progress, and impressions that application developers have with MPI, according to the responses received. We summarize the ...

We provide a status update of ISP, our dynamic formal verifier for MPI programs. ISP determines and explores all relevant schedules of an MPI program. The highlights of this paper are (i) a recap of ISP’s features, and (ii) an overview of work in progress, including a graphical explorer for message passing (GEM) and a distributed MPI analyzer (DMA), an adaptation of ISP for the distributed sett...

.................................................................................................................................... i Declaration ............................................................................................................................... iii Acknowledgements ........................................................................................................

Theoretical models are difficult to apply for the analysis of practical message passing systems used today. We propose a model which can be used for such an analysis. Our framework for message passing is in many ways similar to the framework for transactional database systems. The abstract message passing system is defined in our framework independently of hardware, operating system and means o...