From: Ingo Schmidt <ingo.schmidt@tu...>  20070313 13:53:41

Dear Developers, I've got some questions (still) concerning the NonLinearSolver package of libmesh especially using the SNES part of PETSc. I wrote a test program which just solves the LameNavier equation of motion. The underlying framework is the NewmarkSystem but derivated from the NonLinearImplicitSystem class. During the solution process the Newton iteration always needs only two steps to find the solution and the residual is then still zero. That means there is no iteration. The reasons are the function pointer members of the nonlinear solver class (residual and jacobian) which are called during the iteration from the "__libmesh_petsc_snes_jacobian" or "__libmesh_petsc_snes_residual", respectively. I didn't assigned them.... >:o . a) What's the best way to do this assignment? b) I need two functions, one assembling the "rhs/residual" and one the jacobian, which PETSc is calling permanently during the newton iteration. Both functions are looping over each element of the system. That doubles the numerical effort, or? Roy, you mentioned that glue code is needed to switch to PETSc's SNES package. c) Could you sketch the steps which are necessary? Thanks a lot in advance. Ingo  Dipl.Ing. Ingo Schmidt Institute of Modelling and Computation Hamburg University of Technology Denickestr.17 Building L/ room: 3032 21075 Hamburg http://www.mub.TUHarburg.de 