Dear Shiva, in our group we use MOOCHO, NOX and RYTHMOS in conjunction with deal.II, and there are no issues that I'm aware of.
The only "wrapper" you need, is a "model evaluator" (i.e., a class that implements the interface of EpetraModelEvaluator). Once you have that, you should be good to go, without too much effort. It is my understanding that the wrappers in deal.II are intended to maintain uniformity with the existing LAC section of the library itself, so that you can switch solvers/matrices/vectors from deal.II to Trilinos to PETSc in a almost painless (sometimes...) way, ideally by simply adding a line like using namespace Trilinos; at the top of your include files... However, I strongly advice against wrapping the other Trilinos packages (like NOX, MOOCHO, etc., etc.) inside deal.II. In my experience, the two libraries can coexist without any additional work, as long as you know how to treat vectors/matrices in a unified manner (i.e., through the deal.II wrappers), and have a reasonable knowledge of both libraries. Modifying the Makefile should be enough. Good luck. Luca. -- Luca Heltai <[email protected]> http://people.sissa.it/~heltai/ Scuola Internazionale Superiore di Studi Avanzati Phone: +39 040 3787 449, Office: 732 -- There are no answers, only cross references On Jun 8, 2011, at 12:34 AM, Krishna Siva Shankar Rudraraju wrote: > Hi, > > For some time, I have been solving a nonlinear multi-physics problem using > deal.ii linear solvers at each iteration. But, due to the inherent > nonlinearity, the number of iterations required is significantly high. So, I > am planning to use Trilinos NOX (for its line search and trust region > methods) and Trilinos TEKO (for the multi-physics preconditioners). > > I believe deal.ii doesn't have wrappers for these two specific packages. Is > their any alternative?. > > If their is no alternative, I am considering writing my own wrapper to these > packages (on the lines of deal.ii supplied Trilinos wrappers), and link the > additional Trilinos libraries by modifying the default deal.ii makefile > linking options. Are their any deal.ii specific potential compilation/linking > issues I should be aware of?. > > Thanks, > Shiva Rudraraju > Post-Doc, Computational Physics Group, > University of Michigan Ann Arbor. > > > > > > _______________________________________________ > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
