Michael, > Step-36 uses wrappers to SLEPc (PETSc underlying) and if A can be > formed as a sparse matrix (with a suitable sparsity pattern), then it > appears that the approach taken there would work.
If A is sparse (and large) that may well work. The best thing would be to try it out... :-) > Am I correct in thinking that the idea of avoiding assembling A is a > non-starter, but that either SLEPc or LAPACK are viable options for > finding the eigenvalues of A in deal.II? I don't know about if avoiding assembling A is a good thing or a definate no. There are not enough details about the equation set in your email to decide on this. However I can communicate this: If your A matrix is full or contains alot of non-diagonal entries (or if somehow you explicitly need the inverse of a sparse matrix, which is generally not sparse) do not use SLEPc. SLEPc is a wonderful tool for solving *large* and *sparse* eigenspectrum problems, but SLEPc struggles with *small* and/or *full* eigenspectrum problems; he was not designed for that. Nevertheless, the SLEPcWrappers can handle both the generalized and standard eigenspectrum problems - and that is something! An alternative to consider may be the ARPACK solvers: http://dealii.org/developer/doxygen/deal.II/classArpackSolver.html which uses the Arnoldi eigenspectrum solver (it may be ok to use this for full matrix problems - I do not know the details of the ARPACK solver). The advantage of the ARPACK solvers over the SLEPc solvers is that they use native deal.II objects. Baerbel; any ideas on that last paragraph? Best, Toby ----- Toby D. Young Assistant Professor Institute of Fundamental Technological Problems Polish Academy of Sciences ul Adolfa Pawinskiego 5b 02-106 Warsaw Poland www: http://www.ippt.gov.pl/~tyoung skype: stenografia _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
