Computational & Technology Resources
an online resource for computational,
engineering & technology publications 

CivilComp Proceedings
ISSN 17593433 CCP: 96
PROCEEDINGS OF THE THIRTEENTH INTERNATIONAL CONFERENCE ON CIVIL, STRUCTURAL AND ENVIRONMENTAL ENGINEERING COMPUTING Edited by: B.H.V. Topping and Y. Tsompanakis
Paper 118
Parallel Asynchronous Schwarz Alternating Method for Obstacle Problems on Grid Computing M. Chau^{1}, T. Garcia^{2} and P. Spiteri^{2}
^{1}Advanced Solutions Accelerator, Montpellier, France
M. Chau, T. Garcia, P. Spiteri, "Parallel Asynchronous Schwarz Alternating Method for Obstacle Problems on Grid Computing", in B.H.V. Topping, Y. Tsompanakis, (Editors), "Proceedings of the Thirteenth International Conference on Civil, Structural and Environmental Engineering Computing", CivilComp Press, Stirlingshire, UK, Paper 118, 2011. doi:10.4203/ccp.96.118
Keywords: obstacle problem, parallel algorithm, synchronous method, asynchronous method, Schwarz alternating method, constrained mechanical structures.
Summary
The obstacle problem occurs in many applications such as mechanics, free boundary problems and financial derivatives. In the present study we concentrate mainly on the obstacle problem occurring in civil engineering computing. The problem consists of solving a time dependant nonlinear boundary value problem in which the displacement of the mechanical structure subjected to external strain satisfy an imposed constraint. The previous evolution equation is solved numerically by considering a time marching implicit scheme, where at each time step a stationary nonlinear problem is solved. For each stationary obstacle problem, we consider here a multivalued formulation taking into account the constraint exercised on the displacement.
The Euler implicit time marching scheme has been used for the discretization of the evolution part. Besides, the domain is discretized with a uniform mesh and for the spatial part of the problem, a finite difference method is used. Thus, the numerical solution consists of solving a large linear algebraic system at each time step. So, it is necessary to use parallel iterative methods. The solution of the algebraic system is carried out with the parallel projected Schwarz alternating method; in Schwarz alternating method the subdomains overlap each other. The parallel iterative scheme of computation can be synchronous or asynchronous. In the asynchronous case, computations are carried out in parallel without order nor synchronization, so processors go at their own pace according to their intrinsic characteristics and computational load. The use of an appropriate spatial discretization scheme leads to very interesting properties of the matrices to invert, ensuring the convergence of the parallel iterative synchronous and more generally asynchronous algorithms [1,2]. The algorithms are implemented using Fortran 90 language. The parallelisation is carried out with MPI facilities according to the SPMD paradigm on the Grid'5000 platform, a distributed and heterogeneous architecture. We have used up to 128 machines located in up to two or three distant sites. Results show that all asynchronous schemes of computation scale better than the synchronous ones, as a result of less sensitivity to granularity, network bandwidth, latency and load unbalancing issues. Asynchronous schemes are more adapted to the heterogeneous and distant distributed clusters than the synchronous ones. References
purchase the fulltext of this paper (price £20)
go to the previous paper 
