[petsc-users] SNESNRICHARDSON with LibMesh

[Matthew Knepley]

On Wed, May 27, 2015 at 12:22 PM, Harshad Sahasrabudhe <hsahasra at purdue.edu>
wrote:

> Hi Matt,
>
> Thanks for your comments. I see that SNESNRICHARDSON has changed a lot
> from 3.4.3 to 3.5.3. What do I need to do to make the code compatible with
> both 3.4.3 and 3.5.3?
>

There are two issues:

  1) Making it compatible, which should be no problem since it is just a
SNES. We only changed the implementation.

  2) Comparing results. It may well be that 3.4.3 gives slightly different
results. I would have to look at the exact changes.

  Thanks,

     Matt


> Thanks,
> Harshad
>
> On Tue, May 26, 2015 at 4:25 PM, Matthew Knepley <knepley at gmail.com>
> wrote:
>
>> On Tue, May 26, 2015 at 3:07 PM, Harshad Sahasrabudhe <
>> hsahasra at purdue.edu> wrote:
>>
>>> Hi Matt,
>>>
>>> Actually I just checked, I don't need to re-init the preconditioner
>>> LibMesh solver.
>>>
>>> The wrapper is interfaced to the mesh and makes it easier to assemble
>>> the system matrices. It also keeps track of constraints. The wrapper
>>> function PetscNonlinearSolver::solve updates the system vectors and
>>> matrices, then calls SNESSolve. Thus, I need to call PetscNonlinearSolver::solve
>>> instead of SNESSolve in SNESApplyNPC.
>>>
>>> Do you mean wrap it in SNESSHELL?
>>>
>>
>> Yes, exactly.
>>
>>   Thanks,
>>
>>      Matt
>>
>>
>>> Thanks,
>>> Harshad
>>>
>>> On Tue, May 26, 2015 at 3:25 PM, Matthew Knepley <knepley at gmail.com>
>>> wrote:
>>>
>>>> On Tue, May 26, 2015 at 12:42 PM, Harshad Sahasrabudhe <
>>>> hsahasra at purdue.edu> wrote:
>>>>
>>>>> Hi,
>>>>>
>>>>> I'm using LibMesh for meshing a non-linear equation. I want to use
>>>>> NRichardson non-linear solver to solve the equation. I have a Newton trust
>>>>> region non-linear solver as a pre-conditioner. Both these SNES are created
>>>>> via the LibMesh NonlinearImplicitSystem and PetscNonlinearSolver wrappers.
>>>>>
>>>>> I need to re-init the LibMesh NonlinearImplicitSystem each time the
>>>>> preconditioner is applied. Also, I need the preconditioner to be solved via
>>>>> the LibMesh wrapper instead of directly through SNESSolve. How can make it
>>>>> work this way? Any help is appreciated!
>>>>>
>>>>
>>>> I think the easiest way is to have them both look like SNES objects.
>>>> Then you can just call SNESSetNPC() or use -snes_npc_side left
>>>> -npc_snes_type newtonls
>>>>
>>>> Can you explain what the LibMesh wrapper is doing that you you need? We
>>>> could wrap that object in a SNES, but it sounds
>>>> simpler to me to unwrap it.
>>>>
>>>>   Thanks,
>>>>
>>>>      Matt
>>>>
>>>>
>>>>> Thanks,
>>>>> Harshad
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> What most experimenters take for granted before they begin their
>>>> experiments is infinitely more interesting than any results to which their
>>>> experiments lead.
>>>> -- Norbert Wiener
>>>>
>>>
>>>
>>
>>
>> --
>> What most experimenters take for granted before they begin their
>> experiments is infinitely more interesting than any results to which their
>> experiments lead.
>> -- Norbert Wiener
>>
>
>


-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150527/ef5399bc/attachment-0001.html>