On Wed, May 20, 2015 at 12:06:41PM +0200, Tobias Ritschel wrote: > Just for the record, I found out that Eigen features a number of iterative > solvers though only CG and BiCGSTAB are supported. The others may be found > here <http://eigen.tuxfamily.org/dox/unsupported/namespaceEigen.html>, > together with much other functionality.
My personal (probably biased) opinion is: Do not use it (at least not its sparse iterative solvers) for the following reasons: - It is lacking powerful preconditioners (Multigrid, Domain Decomposition Methods). Which will make it hard to use for real world problems. - While it might be very efficient, sparse linear algebra is somewhat under represented/supported. - It is not parallel (despite small parts using OpenMP), see http://eigen.tuxfamily.org/dox/TopicMultiThreading.html. - Judging from the projects using it, it is not used by many projects doing sparse linear algebra. Like I said I am biased as a DUNE developer (parallel iterative solvers). But if you do not want to use DUNE, then I would rather recommend PETSc, Trilinos or you will find others that might be more suitable from Jack Dongarra's list http://www.netlib.org/utk/people/JackDongarra/la-sw.html Markus -- Dr. Markus Blatt - HPC-Simulation-Software & Services http://www.dr-blatt.de Hans-Bunte-Str. 8-10, 69123 Heidelberg, Germany, USt-Id: DE279960836 Tel.: +49 (0) 160 97590858
signature.asc
Description: Digital signature
_______________________________________________ Opm mailing list Opm@opm-project.org http://www.opm-project.org/mailman/listinfo/opm
