Abstract
Near-optimal CPU efficiency for the basic operations of the Newton-Krylov-BILU iterations for the discontinuous Galerkin method (DGM) can be obtained by exploiting its data locality, as it allows to rewrite these operations in terms of BLAS/LAPACK operations on contiguous vectors and matrices of considerable size. Further significant enhancements are obtained by using single precision pre-conditioners and by exploiting data alignment to improve cache efficiency. The underlying data structures and operations are explained, followed by the comparison of the obtained floating point operation (FLOP) efficiency to the theoretical optimum.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Arioli, M., Duff, I.: Using FGMRES to obtain backward stability in mixed precision. Electronic Transactions on Numerical Analysis 33, 31–44 (2009)
Arnold, D., Brezzi, F., Cockburn, B., et al.: Unified Analysis of Discontinuous Galerkin Methods for Elliptic Problems. SIAM J. Num. Anal. 39, 1749–1779 (2002)
Baboulin, M., Buttari, A., Dongarra, J., et al.: Accelerating Scientific Computations with Mixed Precision Algorithms. Computer Physics Communications (2008) (in press)
Chevaugeon, N., Hillewaert, K., Gallez, X., et al.: Optimal numerical parameterization of discontinuous Galerkin method applied to wave propagation problems. Journal of Computational Physics 223, 188–207 (2007)
Chisholm, T., Zingg, D.: A Jacobian-free Newton-Krylov algorithm for compressible turbulent fluid flows. Journal of Computational Physics 228, 3490–3507 (2009)
Golub, G., Van Loan, C.: Matrix Computations. The Johns Hopkins University Press, Baltimore (1996)
Solin, P., Segeth, K., Dolezel, I.: Higher Order Finite Element Methods. Studies in Advanced Mathematics. Chapman and Hall/CRC (2004)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2010 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Hillewaert, K. (2010). Exploiting Data Locality in the DGM Discretisation for Optimal Efficiency. In: Kroll, N., Bieler, H., Deconinck, H., Couaillier, V., van der Ven, H., Sørensen, K. (eds) ADIGMA - A European Initiative on the Development of Adaptive Higher-Order Variational Methods for Aerospace Applications. Notes on Numerical Fluid Mechanics and Multidisciplinary Design, vol 113. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-03707-8_2
Download citation
DOI: https://doi.org/10.1007/978-3-642-03707-8_2
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-03706-1
Online ISBN: 978-3-642-03707-8
eBook Packages: EngineeringEngineering (R0)