On 23/03/2009, John Peterson wrote: > On Mon, Mar 23, 2009 at 2:08 PM, Kirk, Benjamin (JSC-EG311) > <[email protected]> wrote: >> John - since you have at least used SLEPc, would you please give >> this patch >> a look? I think it is great that they did this for us!! > > The patch is now checked in. At the moment I could only easily test > against Petsc/Slepc 2.3.1, but whenever I get around to installing > Petsc 3.0.0 I will take a look at this again. I had to make a few > changes to allow backwards-compatibility since EPSKRYLOVSCHUR was not > yet available in Slepc 2.3.1. Since I don't know when exactly it > showed up, you now only get the Krylov-Schur solver when Petsc/Slepc > 3.0.0 is available, otherwise you get ARNOLDI, which was the old > default behavior. > > Perhaps the slepc developers could suggest a different or better > default for versions of slepc prior to the introduction of > Krylov-Schur? > > Thanks, > -- > John
The Krylov-Schur solver was introduced in SLEPc 2.3.2, see http://www.grycap.upv.es/slepc/download/changes.htm I would suggest to set EPSKRYLOVSCHUR whenever it is available. An alternative is to let SLEPc set the default solver. The Krylov-Schur method is mathematically equivalent to implicitly restarted Arnoldi, the method of Arpack, so there is currently no point in using SLEPc with Arpack. Thanks, Jose ------------------------------------------------------------------------------ _______________________________________________ Libmesh-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-devel
