Distributed algebraic multigrid for finite element computations

Authors:
C. Farris;M. Misra
Affiliations:
Department of Mathematical and Computer Sciences Colorado School of Mines, Golden, CO 80401, U.S.A.;Department of Mathematical and Computer Sciences Colorado School of Mines, Golden, CO 80401, U.S.A.
Venue:
Mathematical and Computer Modelling: An International Journal
Year:
1998

Citing 8
Cited 0

Parallel adaptive full-multigrid methods on message-based multiprocessors

Parallel Computing
Algebraic multigrid theory: The symmetric case

Applied Mathematics and Computation - Second Copper Mountain conference on Multigrid methods Copper Mountain, Colorado
A multigrid method for multiprocessors

Applied Mathematics and Computation - Second Copper Mountain conference on Multigrid methods Copper Mountain, Colorado
Two multigrid methods for three-dimensional problems with discontinuous and anisotropic coefficients

SIAM Journal on Scientific and Statistical Computing
Programming the finite element method (2nd ed.)

Programming the finite element method (2nd ed.)
An adaptive algebraic multigrid for reactor criticality calculations

SIAM Journal on Scientific Computing
Concepts for efficient multigrid implementation on SUPRENUM-like architectures

Parallel Computing
On the parallelization of an adaptive multigrid algorithm for a class of flow problems

Parallel Computing

Quantified Score

Hi-index	0.98

Visualization

Abstract

The Finite Element Method has been successfully applied to a variety of problems in engineering, medicine, biology, and physics. However, this method can be computationally intensive, particularly for problems in which an unstructured mesh of elements is generated. In such situations, the Algebraic Multigrid (AMG) can prove to be a robust method for solving the discretized linear systems that emerge from the problem. Unfortunately, AMG requires a large amount of storage (thus causing swapping on most sequential machines), and typically converges slowly. We show that distributing the algorithm across a cluster of workstations can help alleviate these problems. The distributed algorithm is run on a number of geomechanics problems that are solved using finite elements. The results show that distributed processing is extremely useful in maintaining the performance of the AMG algorithm with increasing problem size, particularly by reducing the amount of disk swapping required.