Paraytec (United Kingdom)

PC. Macey (I) N.K. Allsopp (2) AS. Gill (2) (I) PAFEC Ltd. Strelley Hall. Nottingham, N08 6PE (2) Parallel Applications Centre, 2 Venture Road, Chilworth, Sothampton. Stages 2 and 3 ofthe solution process are very ef ciently parallelized. These stages involve little communication between processors. Stage 1 is slower when run on more than 1 processor. This is because there is a relatively high ratio ofcommunication to computation Stage 4 suffers from the same problem to a lesser extent. However as the problemsize increases the ratio of communication to computation decreases for all stages. It is thus helievcd that the viability of paralch vibroacoustic analysis has been shown for large problems. Work is in progress on some larger problems which will hopefully con rm this conclusion.

The Navier-Stokes Solver TFS developed by the Institute of Aerodynamics of the RWTH Aachen University is currently used in a multidisciplinary project to simulate the air flow through the human nose. In order to reduce the runtime of the expensive computations the ParaWise/CAPO automatic parallelization environment has been used to assist in the nested OpenMP parallelization of the TFS multi-block code targeted at Sun Microsystems shared memory parallel systems. Further manual tuning improved the scalability of the OpenMP approach.KeywordsParallel VersionParallel ExecutionLoad ImbalanceParallel RegionCache CoherenceThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

One of the goals of the HPC-Europa project is to provide users with a Single Point of Access (SPA) to the resources of HPC centers in Europe. To this end, the HPC-Europa Portal is being built to provide transparent, uniform, flexible and intuitive user access to HPC-Europa resources. In this paper, we present a mechanism that enables end-users to transparently access the diverse services available in the HPC-Europa environment. The uniform job submission interface that uses this mechanism, utilizing the job specification description language (JSDL), is described. We also present the architecture of the SPA, based on the GridSphere portal framework. Finally, we discuss the various interoperability problems encountered, in particular those concerning job submission, security and accounting.

Data sets being managed in grid environments today are growing at a rapid rate, expected to reach 100s of petabytes in the near future. Managing such large data sets poses challenges for efficient data access, data publication and data discovery. In this paper we focus on the data publication and discovery process through the use of descriptive metadata. This metadata describe the properties of individual data items and collections. We discuss issues of metadata services in service rich environments, such as the grid. We describe the requirements and the architecture for such services in the context of grid and the available grid services. We present a data model that can capture the complexity of the data publication and discovery process. Based on that model we identify a set of interfaces and operations that need to be provided to support metadata management. We present a particular implementation of a grid metadata service, basing it on existing grid services technologies. Finally we examine alternative implementations of that service.

Distributed shared memory has been recognized as an alternative programming model to exploit the parallelism in distributed memory systems since it provides a higher level of abstraction than simple message passing. DSM combines the simple programming model of shared-memory with the scalability of distributed memory machines. This paper presents DSMPI, a parallel library that runs atop of MPI and provides a distributed shared memory abstraction. It provides an easy-to-use programming interface, is flexible, portable and supports heterogeneity. Moreover, it supports different coherence protocols and models of consistency. We present some performance results taken in a network of workstations and in a Cray T3D which show that DSMPI can be competitive with MPI for some applications.