Implicit domain decomposition scheme for parallel dynamic finite element geotechnical analysis

Authors:
Zhao Ben;Goh Siang Huat;Lee Fook Hou
Affiliations:
National University of Singapore, Department of Civil Engineering, Singapore;National University of Singapore, Department of Civil Engineering, Singapore;National University of Singapore, Department of Civil Engineering, Singapore
Venue:
SEPADS'10 Proceedings of the 9th WSEAS international conference on Software engineering, parallel and distributed systems
Year:
2010

Citing 4
Cited 0

Solution of sparse positive definite systems on a shared-memory multiprocessor

International Journal of Parallel Programming
An improved spectral graph partitioning algorithm for mapping parallel computations

SIAM Journal on Scientific Computing
Parallel optimisation algorithms for multilevel mesh partitioning

Parallel Computing - Special issue on graph partioning and parallel computing
A distributed memory parallel element-by-element scheme based on Jacobi-conditioned conjugate gradient for 3D finite element analysis

Finite Elements in Analysis and Design

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents a parallel finite element method for large-scale analysis of seismic effects on geotechnical systems. A parallel form of the element-by-element modified Jacobi preconditioned conjugate gradient method is used to solve the equation system. An implicit domain decomposition scheme is proposed which divides the finite element domain into sub-domains based directly on element numbering. This obviates a lot of the graphical difficults associated with domain partitioning; at the same time, it enhances automated decomposition and balanced workload distribution. The parallel method is implemented using a message passing programming model on a PC cluster connected with high-speed network. Two numerical examples are carried out to examine the performance of the proposed decomposition scheme. The results shows that, using this parallelization method, it is possible to analyze large-scale geotechnical models, involving 4,000,000 degrees of freedom or more, with a reasonable time span on a PC cluster.