If you have superlu installed, I've found -pc_type bjacobi -sub_pc_type ilu -sub_pc_mat_solver_package superlu to be much better than the standard ILU. Note needs to be PETSc 3.3-p2 at least because there was a bug where superlu only got used in the first iteration.
On Thu, Sep 13, 2012 at 3:30 PM, Roy Stogner <royst...@ices.utexas.edu>wrote: > > My favorite preconditioner has always been ILU4 (wrapped in block > Jacobi in parallel), and that's worked well enough that except for a > couple applications I've never bothered trying to do better. > > Now I'm discovering that ILU4 fails horribly on our little old > vector_fe_ex3 (curl-curl) example. I have no idea why one step of ILU > converges in a hundred-odd iterations and four steps doesn't make it > to square root of tolerance after 5000 iterations, but clearly I need > new LIBMESH_OPTIONS. > > Any suggestions? > --- > Roy > > > ------------------------------------------------------------------------------ > Live Security Virtual Conference > Exclusive live event will cover all the ways today's security and > threat landscape has changed and how IT managers can respond. Discussions > will include endpoint security, mobile security and the latest in malware > threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ > _______________________________________________ > Libmesh-users mailing list > Libmesh-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/libmesh-users > ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
