A Stochastic-based Optimized Schwarz Method for the Gravimetry Equations on GPU Clusters
Ecole Centrale Paris, France
21st International Conference on Domain Decomposition Methods, 2013
@article{ahamed2013stochastic,
title={A Stochastic-based Optimized Schwarz Method for the Gravimetry Equations on GPU Clusters},
author={Ahamed, Abal-Kassim Cheik and Magoules, Fr{‘e}d{‘e}ric},
year={2013}
}
By giving another way to see beneath the Earth, gravimetry refines geophysical exploration. In this paper, we evaluate the gravimetry field in the Chicxulub crater area located in between the Yucatan region and the Gulf of Mexico which shows strong gravimetry and magnetic anomalies. High order finite elements analysis is considered with input data arising from real measurements. The linear system is then solved with a domain decomposition method, namely the optimized Schwarz method. The principle of this method is to decompose the computational domain into smaller subdomains and to solve the equations on each subdomain. Each subdomain could easily be allocated to one single processor (i.e. the CPU), each iteration of the optimized Schwarz method involving the solution of the equations on each subdomain (on the GPU). Unfortunately, to obtain high speed-up, several tunings and adaptations of the algorithm should be carrefully performed, such as data transfers between CPU and GPU, and data structures, as described in [3, 2].
September 2, 2013 by hgpu