Parallel Computational Fluid Dynamics 1993. New Trends and by J. Hauser, P. Leca, A. Ecer, Jacques Periaux

By J. Hauser, P. Leca, A. Ecer, Jacques Periaux

This quantity comprises the papers awarded on the Parallel Computing Fluid Dynamics '93 convention, Paris, 1993. a variety of issues are lined together with: networked desktops, info parallel programming, area decomposition, Euler and Navier-Stokes solvers.

Researchers during this sector will locate this quantity an invaluable reference during this swiftly constructing box

Show description

Read Online or Download Parallel Computational Fluid Dynamics 1993. New Trends and Advances PDF

Best data modeling & design books

Modular Ontologies: Concepts, Theories and Techniques for Knowledge Modularization

This booklet constitutes a set of analysis achievements mature adequate to supply a company and trustworthy foundation on modular ontologies. It provides the reader an in depth research of the state-of-the-art of the study sector and discusses the new thoughts, theories and methods for wisdom modularization.

Advances in Object-Oriented Data Modeling

Until eventually lately, info structures were designed round diversified enterprise services, akin to debts payable and stock regulate. Object-oriented modeling, against this, buildings platforms round the data--the objects--that make up a number of the enterprise capabilities. simply because information regarding a specific functionality is proscribed to 1 place--to the object--the process is protected against the consequences of swap.

Introduction To Database Management System

Designed particularly for a unmarried semester, first direction on database platforms, there are four facets that differentiate our e-book from the remaining. simplicity - as a rule, the know-how of database structures may be very obscure. There are

Additional info for Parallel Computational Fluid Dynamics 1993. New Trends and Advances

Example text

All rights reserved. uk Abstract We review the published achievements in the use of parallel processing for environmental modelling projects. Results from machines with a range of parallel architectures show that in many cases performance is limited to around 5%-20% of the peak performance, mainly by the computational rate that can be achieved on a single processor, and to a lesser extent by interprocessor communications and by the problems of load balancing. 1. INTRODUCTION Computer modelling is becoming more and more important to environmental science.

47 In most cases there was no problem in obtaining a good parallel efficiency provided that a sufficiently large problem was used. The model run by McBryan achieves particularly impressive results with a parallel efficiency of 96% from 1 to 128 nodes. This high efficiency may well be because the large amount of computational work required at each grid point is highly effective at masking the parallel overheads introduced by communications. Table 1. Summary of the best parallel speed-up achieved for an environmental modelling application from a number of references.

I i J2F .................................... • //" 100 0 I i I i t 1 I 1 4 8 12 16 20 24 28 32 Number of processors Figure 1. Performance in Mflop/s of the POLMP code on the Intel iPSC/860 as a function of the number of processors used. The dashed line shows the ideal performance obtained by scaling the single node performance. Each processor works on a sea area of 128 x 64 points with 16 vertical modes and the model was run for 200 timesteps. 43 2. REVIEW OF RESULTS FROM THE LITERATURE The Proudman Oceanographic Laboratory Multiprocessing Program (POLMP) (Ashworth and Davies, 1992, Ashworth, 1992, 1993) is a demonstration three-dimensional shallow sea model which has been specifically designed to run efficiently on computers with a wide range of parallel architectures.

Download PDF sample

Rated 4.91 of 5 – based on 15 votes