CERFACS CERFACS Scientific Activity Report Jan. 2010 â Dec. 2011
CERFACS CERFACS Scientific Activity Report Jan. 2010 â Dec. 2011
CERFACS CERFACS Scientific Activity Report Jan. 2010 â Dec. 2011
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
PARALLEL ALGORITHMS PROJECT<br />
We are currently working on this method and some variations as the block implementation or its use to<br />
compute the harmonic Ritz vectors (G.W. Stewart, SIAM J. Matrix Anal. Appl., 2002).<br />
This work has been presented in conferences (e.g., P. Salas, L. Giraud, J. Muller, G. Staffelbach, T. Poinsot,<br />
”Stability of azimuthal modes in annular combustion chambers”, INCA, November <strong>2011</strong>) and a paper has<br />
been accepted for its publication in the journal of Combustion and Flame (<strong>CERFACS</strong> Technical <strong>Report</strong>,<br />
TR/CFD/11/35, J.-F. Parmentier, P. Salas, P. Wolf, G. Staffelbach, F. Nicoud and T. Poinsot, ”A simple<br />
analytical model to study and control azimuthal instabilities in annular combustion chambers”).<br />
5.10 Minimizing the backward error in the energy norm with<br />
conjugate gradients.<br />
S. Gratton : INPT-IRIT, UNIVERSITY OF TOULOUSE AND ENSEEIHT, France ;<br />
P. Jiránek : TECHNICAL UNIVERSITY OF LIBEREC AND <strong>CERFACS</strong>, Czech Republic and France ;<br />
X. Vasseur : <strong>CERFACS</strong>, France<br />
In [ALG57], we derive backward error formulas for a linear system of equations in general norms induced<br />
by given symmetric positive definite matrices and consider a special case of a backward error induced by<br />
the energy norm when the system matrix is symmetric positive definite. We study the convergence of the<br />
conjugate gradient method (CG) with respect to this energy backward error. For that purpose we construct a<br />
hypothetical variant of CG called CGBACK which constructs the approximations that actually minimize the<br />
energy backward error over the associated Krylov subspaces and can therefore be considered as an analog<br />
of the GMBACK/MINPERT algorithms of Kasenally and Simoncini. We show that the optimal CGBACK<br />
approximation is a scalar multiple of the current CG approximation with the coefficient depending only<br />
on the weighting parameters appearing in the definition of the backward error and on the relative energy<br />
norm of the error of the current CG iterate. In addition when CG makes a moderate progress in terms of the<br />
energy norm of the error then the energy backward errors of the subsequent CG approximations start to be<br />
very close to the optimal energy backward errors of CGBACK approximations. In this way we deduce that<br />
CG approximations almost minimize the energy backward error.<br />
5.11 Adaptive version of simpler GMRES.<br />
P. Jiránek : TECHNICAL UNIVERSITY OF LIBEREC AND <strong>CERFACS</strong>, Czech Republic and France ;<br />
M. Rozložník : ACADEMY OF SCIENCES OF THE CZECH REPUBLIC, Czech Republic<br />
In [ALG29], we propose a stable variant of Simpler GMRES by Walker and Zhou (1994). It is based on<br />
the adaptive choice of the Krylov subspace basis at given iteration step using the intermediate residual<br />
norm decrease criterion. The new direction vector is chosen as in the original implementation of Simpler<br />
GMRES or it is equal the normalized residual vector as in the GCR method. We show that such adaptive<br />
strategy leads to a well-conditioned basis of the Krylov subspace and we support our theoretical results with<br />
illustrative numerical examples.<br />
<strong>CERFACS</strong> ACTIVITY REPORT 19