Re: [Libmesh-users] SolverConfiguration?

[Derek G. <fri...@gm...>]

Ok - I just thought it was kind of odd that it has some data members and
stuff - but then doesn't do anything by default.

We use the command-line options for PETSc... but we set them
programmatically.  Gives us the most flexibility because there are
thousands of them and they are always changing.

Thanks for the reply - I was just making sure I wasn't missing anything.

Derek

On Fri, Apr 27, 2018 at 7:21 PM David Knezevic <dav...@ak...>
wrote:

> On Fri, Apr 27, 2018 at 6:40 PM, Derek Gaston <fri...@gm...> wrote:
>
>> What is up with SolverConfiguration?  It's abstract... and I can't see a
>> single implementation of it anywhere in libMesh.  In addition - everything
>> is public.
>>
>> Shouldn't there be a few implementations of this (like for PETSc, etc.)?
>>
>> What am I missing?
>>
>
>
> You're not missing anything, there are no subclasses of
> SolverConfiguration in libMesh itself. It is demonstrated in
> systems_of_equations_ex6, though.
>
> Personally, I use it a lot in app codes. I find it to be a good way to
> control solver options programmatically, i.e. without relying on command
> line options. You can override configure_solver and basically write
> whatever PETSc code you want in there to set the solver options.
>
> We could certainly add SolverConfiguration implementations to libMesh, but
> the main point was to allow users to customize solver options themselves,
> so I guess no one had a need to add any specific cases to libMesh so far.
>
> David
>