Storming Media: Pentagon Reports and DocumentsPentagon Reports: Fast. Definitive. Complete.     
New Account »
Forgot Password?
Advanced Search »
Math and StatisticsNumerical Mathematics

An Efficient Parallel Finite-Element-Based Domain Decomposition Iterative Technique With Polynomial Preconditioning

Authors: Yu Liang; Ramdev Kanapady; Kumar Tamma; MINNESOTA UNIV MINNEAPOLIS DEPT OF COMPUTER SCIENCE
Abstract:
An efficient parallel finite element-based domain decomposition iterative technique with polynomial preconditioning with particular attention to the GMRES solver is presented. Unlike the standard row-oriented partitioning of a matrix, finite element based domain decomposition with polynomial preconditioning circumvents the assembly of matrix, reordering of matrix, redundant computations associated with the interface elements, numerical problems associated with local preconditioner, and costly global preconditioner construction. A dramatic reduction in parallel overhead both in terms of computation and communication results in a highly scalable solver. The parallel performance results for large-scale static and dynamic problems on the IBM SP2 and the SGI Origin are presented.

Limitations: APPROVED FOR PUBLIC RELEASE
Description: Technical rept.
Pages: 29
Report Date: 18 JAN 2005
Contract Number: DAAH0496C0086, DAAD190120014
Report Number: A784934
Keywords relating to this report:
*FINITE ELEMENT ANALYSIS
*PARALLEL PROCESSING
*POLYNOMIALS
COMMUNICATION AND RADIO SYSTEMS
COMPUTATIONS
DYNAMICS
INTERFACES
REDUNDANCY
STATICS
Email This Abstract