Published 1986 | Version v1
Book

Parallel computation of multiple-scale problem

  • 1. Lawrence Livermore National Lab., P.O. Box 808, Livermore, CA

Description

The efficient use of multiprocessor computers requires the introduction of new numerical algorithms, and the identification by asymptotic analysis of concurrencies inherent in the governing equations of multiple-scale problems is a good way to develop such algorithms. The language used must be able to maintain these concurrencies. Finally, depending on the problem and the computer, this concurrency may not be sufficient to keep all the processors busy, so that the language must be able to identify additional concurrencies. These ideas are illustrated in an example based on a convection-diffusion equation

Additional details

Publishing Information

Publisher
Society for Industrial and Applied Mathematics.
Imprint Place
Philadelphia, PA (USA)
ISBN
0-89871-201-7
Imprint Title
New computing environments:Parallel, vector and systolic
Journal Page Range
p. 136-153.

Conference

Title
parallel, vector and systolic.
Acronym
ARO workshop on new computing environments
Dates
7-9 Nov 1984.
Place
Stanford, CA (USA).

INIS

Country of Publication
United States
Country of Input or Organization
United States
INIS RN
18029239
Subject category
S99: GENERAL AND MISCELLANEOUS; S71: CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS;
Resource subtype / Literary indicator
Conference
Descriptors DEI
ALGORITHMS; ARRAY PROCESSORS; ASYMPTOTIC SOLUTIONS; CONVECTION; DIFFUSION; EFFICIENCY; EQUATIONS; PARALLEL PROCESSING; PERFORMANCE; PROGRAMMING LANGUAGES
Descriptors DEC
COMPUTERS; DIGITAL COMPUTERS; ENERGY TRANSFER; HEAT TRANSFER; PROGRAMMING