Hello, Derek, Roy told me that you are using JFNK now and keep the sparse matrix to fill with an inexact Jacobian for preconditioning, I am interesting about this too, so I want to know how to implement JFNK with nonlinearImplicitSystem? Just running with the line command "_ksp_matrix_free" or "_ksp_mf_operator"?
I also want to use a matrix_free preconditioner and I need to store all diagonal dense matrix of each elem for speed up the user-provided "Kx=y" process, can you give me some suggestions? Do I need to write a matrix_free system inherited from ExplicitSystem or NonlinearImplicitSystem? Or just modify the solve() of petsc_nonlinear_solver.C? For my discontinuous galerkin code, because the elem_based FE family, I need to compute the sparse pattern with command "--implicit_neighbor_dofs", if I just need the diagonal dense matrix, how to implement? Just without this command? Best regards! Luyi -- Gas Turbine Research Center email: [EMAIL PROTECTED] ------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/ _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
