scientific article; zbMATH DE number 1460607

From MaRDI portal
Publication:4485696





The articles of this volume will not be indexed individually. From the introduction: When compiling a proceedings volume subsequent to a research symposium, editors are faced with the question: ``Are we able to provide the reader with information which is still current state-of-the-art and will remain relevant for research in the following years? In fast developing areas such as High Performance Computing and Networking (HPCN), this is often a very difficult question to answer. However, I believe that the timing of the 8th NEC Research Symposium, held in Berlin in May 1997, was most fortunate and enables me to give a most positive answer to the above question. The topic of the Symposium (``Heterogeneous Computing and Multi-Disciplinary Applications) addressed, for both computing platforms and scientific simulations, the expansion from what was, and still is currently, normal practice to new approaches, possibilities and related problems. The timing of the symposium was fortunate as we were able to bring together established successes with areas where, in general, the problem definition was clear but only first steps towards its solution had been made. The lively discussions which were a most rewarding feature of the symposium, in particular during the panel discussion ``Meta-Computers: Resource sharing or Grand-Challenge Tool?, centred around questions which remain topics of intense activity today. Clustered computing: should we be looking for commodity PC-clusters or more efficient ways to exploit the networking of large computing centres? As the expectations of numerical simulation (for academic research and for use in industrial design) increase, what strategies should be adopted to handle multi-physics, multiscale applications? Once we have the heterogeneous computing systems with interacting applications software, what technologies are required to incorporate these within the industrial concern's design infrastructure? The proceedings of the 8th NEC Research Symposium won't answer all of these questions. What this volume provides is a demonstration that major advances are being made towards providing the technological solutions to these questions as well as posing problems to be addressed in the future. The development and scientific use of large scale parallel machines is complemented by the scheduling of applications running on heterogeneous, multi-site platforms. Approaches for combining parallel application codes to solve fluid-structure and fluid-fluid interactions are complemented by new techniques for simulation on multi-site platforms, post-processing of data from large-scale HPC applications and a look-ahead to the industry of the future using concurrent engineering for virtual design.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4485696)