Re: [deal.II] Controllins fill-in, abs. tol and rel. tol or smoother and coarse solver of TrilinosWrappers::PreconditionAMG

[Wolfgang Bangerth]

On 3/8/19 9:26 AM, Bruno Blais wrote:
> 
> Clearly, parameters like fill-level, rel. treshold and abs. threshold could 
> be 
> modified. Is there any way to access them and modify them? I have not been 
> able to as of now.

Bruno -- Jean-Paul already outlined what's necessary. But I wanted to add that 
it would be really interesting if you could post what you find once you make 
this sort of experiment. I don't think that we've made systematic studies on 
what works best in the context of our discretizations, and it would be rather 
useful to hear what you find!

Best
  W.

-- 

Wolfgang Bangerth  email: bange...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [deal.II] projecting boundary values for H(div) [and maybe H(curl)]

[Wolfgang Bangerth]

On 3/10/19 4:09 AM, Konrad wrote:
> 
> just a quick question. Is there a special reason why for example the function 
> VectorTools::project_boundary_values_div_conforming is restricted to 
> Raviart-Thomas elements? The principle should be the same for example for 
> Brezzi-Douglas-Marini elements, right? Or am I missing something here?

Correct. I suspect it's because nobody has tried it for the BDM element and 
thought through whether the details are correct there as well.

What happens if you just try it? You might even be able to just take some of 
the tests under the tests/ directory and copy/adjust them to use the BDM 
element to see what happens. If it doesn't work, I suspect that the 
adjustments will be minor. We'd of course take any kind of patch you come up 
with!

Best
  W.


-- 

Wolfgang Bangerth  email: bange...@colostate.edu
www: http://www.math.colostate.edu/~bangerth/

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [deal.II] Controllins fill-in, abs. tol and rel. tol or smoother and coarse solver of TrilinosWrappers::PreconditionAMG

[Jean-Paul Pelteret]

Dear Bruno,

Indeed, it appears that one isn't able to change the settings passed to the AMG 
preconditioner other than those set by the AdditionalData object. Your only 
current alternative is to create the Teuchos::ParameterList yourself, which you 
can send in to the call to initialize() instead of the AdditionalData object. 

This is, in general, not an optimal solution, as choosing the correct settings 
for the preconditioner is not trivial. So I’ve opened this PR to (hopefully) 
give the solution that you were hoping for.
https://github.com/dealii/dealii/pull/7793 


Best,
Jean-Paul

> On 08 Mar 2019, at 17:26, Bruno Blais  wrote:
> 
> Hello everyone,
> Just a small question.
> Looking at the documentation and at the .h of the  
> TrilinosWrappers::PreconditionAMG  I have not found a way to access the 
> smoother and coarse solver to play with their parameters. Say if you use an 
> ILU or ILUT smoother.
> When you use the verbose output of the solver you get status report such as:
> 
> Smoother (level 2) : IFPACK, type=`ILU',
> Smoother (level 2) : post,overlap=1
> Smoother (level 2) : level-of-fill=0,rel. threshold=1,abs. threshold=0
> Smoother (level 2) : Setup time : 0.00463979 (s)
> 
> Coarse solve (level 3) : IFPACK, type=`ILU',
> Coarse solve (level 3) : post,overlap=1
> Coarse solve (level 3) : level-of-fill=0,rel. threshold=1,abs. threshold=0
> 
> 
> Clearly, parameters like fill-level, rel. treshold and abs. threshold could 
> be modified. Is there any way to access them and modify them? I have not been 
> able to as of now.
> 
> I would prefer to keep using the Wrappers as I find their usage to be 
> relatively easy and straightforward.
> Thank you for everything,
> Best
> Bruno
> 
> -- 
> The deal.II project is located at http://www.dealii.org/ 
> 
> For mailing list/forum options, see 
> https://groups.google.com/d/forum/dealii?hl=en 
> 
> --- 
> You received this message because you are subscribed to the Google Groups 
> "deal.II User Group" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to dealii+unsubscr...@googlegroups.com 
> .
> For more options, visit https://groups.google.com/d/optout 
> .

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[deal.II] How to apply first kind of boundary conditions to one component of a vector

[chenxi . edu1997]

Hi, everyone.
I am trying to solve a coupling geomechanics and fluid problem using the 
space:(dim = 2 )

> fe (FE_RaviartThomas(degree), 1,
>
> //FE_Q(degree+1),dim,
>
> FE_DGQ(degree), 1,
>
> FE_DGQ(degree), 1,
>
> FE_Q(degree+1), dim),
>
> and 

> FE_Q(degree+1), dim)
>
>
represent the displacement of the rock.

In order to check if this code is right, I want to use a model with 
analytical solution to verify it. And I need to make the displacement in y 
direction to be zero.
At first I use fe.component_mask(displacement), 
But this works for both displacements in x and y directions.

Then I try to construct a ComponentMask vector:
I make the 2*dim+1 term to be false(which represent the constraints of y 
direction)

However, I got this error:

>  void 
>> dealii::VectorTools::{anonymous}::do_interpolate_boundary_values(const 
>> M_or_MC&, const DoFHandlerType&, const 
>> std::map*>&, 
>> std::map&, const dealii::ComponentMask&) [with int 
>> dim = 2; int spacedim = 2; number = double; DoFHandlerType = 
>> dealii::DoFHandler; M_or_MC = dealii::Mapping]
>
> The violated condition was: 
>
> cell->get_fe().is_primitive (i)
>
> Additional information: 
>
> This function can only deal with requested boundary values that 
>> correspond to primitive (scalar) base elements
>
>
>
Is there a good way to solve this problem?
Thank you in advance!


Best,
Chen 

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[deal.II] projecting boundary values for H(div) [and maybe H(curl)]

[Konrad]

Hi all,

just a quick question. Is there a special reason why for example the 
function VectorTools::project_boundary_values_div_conforming is restricted 
to Raviart-Thomas elements? The principle should be the same for example 
for Brezzi-Douglas-Marini elements, right? Or am I missing something here?

Konrad

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: [deal.II] Nedelec elements of second kind and Brezzi-Douglas-Marini spaces

[Konrad]

>
> I must admit that I don't have a standard reference for it. There are a 
> number of books on high order and hp finite element methods that often 
> provide a pretty good overview of the families of finite elements out 
> there. Maybe the one by Solin, Segeth, Dolezel is a good starting point?  
>

Thank you, I will see if I can make it. 
Konrad

-- 
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see 
https://groups.google.com/d/forum/dealii?hl=en
--- 
You received this message because you are subscribed to the Google Groups 
"deal.II User Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to dealii+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.