Hi Roy, I was able to spot where the value is being printed. Let me see if I can control it. Also, Is there any way to prevent re-integration of the jacobian on every non-linear step? One possible way is to just store the matrix and pass the matrix to the jacobian function. Thanks, Subramanya
----- Original Message ----- From: "Roy Stogner" <[email protected]> To: "Subramanya Gautam Sadasiva" <[email protected]> Cc: [email protected] Sent: Thursday, August 23, 2012 9:00:23 AM Subject: Re: [Libmesh-users] Output from nonlinear solver. On Thu, 23 Aug 2012, Subramanya Gautam Sadasiva wrote: > I am deriving from the TransientNonlinearImplicitSystem,and then I > just connect to a nonlinear solver to which I supply jacobian, > residual and bounds. I guess it eventually goes through to the > petscnonlinear solver class though I don't directly use that class > anywhere. Subramanya Yeah - if you're using the defaults for a NonlinearSystem with PETSc installed, then you're probably getting output printed via __libmesh_petsc_snes_monitor in petsc_nonlinear_solver.C. We don't seem to have any way to turn that off from user code. If you wanted to add a way, we'd happily accept a patch. Best way would probably be to put a "bool quiet;" in the NonlinearSolver base class, cast ctx to a PetscNonlinearSolver in the __libmesh_petsc_snes_monitor, and test PetscNonlinearSolver::quiet before printing. --- 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 [email protected] https://lists.sourceforge.net/lists/listinfo/libmesh-users
