[petsc-users] Near nullspace for preconditioners other than GAMG?

[Mark Adams]

Penalty methods are hell on solvers.  There is a good change that GAMG will
do the right thing with a finite threshold. and keep this stiff section
separate.  You might want to start wit a small number of levels and a large
coarse grid and get it working.  Then add more levels and see if it breaks.

On Sun, Mar 8, 2015 at 11:38 AM, David Knezevic <david.knezevic at akselos.com>
wrote:

> On Sun, Mar 8, 2015 at 10:31 AM, Mark Adams <mfadams at lbl.gov> wrote:
>
>>
>>
>>
>>> Another situation I've been meaning to ask about: If the elasticity
>>> model includes some "near rigid" regions (very high Young's modulus)
>>>
>>
>> Do you mean Poisson ratio?
>>
>
>
> No, just high Young's modulus in subregions. Engineers sometimes implement
> a rotation boundary condition about a node by including rigid (or, in
> practice, very high Young's modulus) parts in a model. For example, imagine
> including a rigid pyramid in a model, in which the tip of the pyramid is is
> clamped so that the base of the pyramid (which is connected to a surface in
> the model) can rotate about the tip.
>
> This works fine with a direct solver, so I was curious about getting a
> good iterative solver for this case too.
>
>
>
>> Nearly incompressible is hard.  bigger smoothers, like ASM, can help.
>>
>
> Agreed. But the Poisson ratio is still just 0.3 or so in this context, so
> locking and near incompressibility aren't an issue here.
>
> Thanks,
> David
>
>
>
>
>
>> On Sat, Mar 7, 2015 at 5:57 PM, Mark Adams <mfadams at lbl.gov> wrote:
>>
>>> FYI, stüben used classical AMG for elasticity but he has articulated his
>>> code for elasticity more than Hypre as I understand it.  Hypre can work OK
>>> for elasticity in my experience.  Its worth a try.
>>>
>>> Mark
>>>
>>> On Thu, Mar 5, 2015 at 5:27 PM, David Knezevic <
>>> david.knezevic at akselos.com> wrote:
>>>
>>>> OK, got it, thanks!
>>>>
>>>> David
>>>>
>>>>
>>>> On Thu, Mar 5, 2015 at 5:08 PM, Jed Brown <jed at jedbrown.org> wrote:
>>>>
>>>>> David Knezevic <david.knezevic at akselos.com> writes:
>>>>> > I was just wondering if its possible to achieve the same sort of
>>>>> thing with
>>>>> > other AMG solvers (e.g. BoomerAMG)? I assume that
>>>>> MatSetNearNullSpace does
>>>>> > nothing for external solvers like hypre, right?
>>>>>
>>>>> It is used by ML (smoothed aggregation), but not BoomerAMG (classical
>>>>> AMG) which uses an algorithm that doesn't have a natural place for such
>>>>> information.  To my knowledge, classical AMG is not widely used for
>>>>> elasticity.  It is very robust for M-matrices.
>>>>>
>>>>
>>>>
>>>
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-users/attachments/20150310/507d0c46/attachment.html>