[deal.II] Is parallel direct solver extremely slow?

Hi all, I learn form Prof. Wolfgang's lectures that parallel direct solver is really competitive, when compared with the parallel distributed solver. So I try to modify the solver in step-17 into a parallel direct solver, and I found it *extremely solw*. The following is the details how I mo

[deal.II] deal.II Newsletter #46

Hello everyone! This is deal.II newsletter #46. It automatically reports recently merged features and discussions about the deal.II finite element library. ## Below you find a list of recently proposed or merged features: #7131: Doc fix: Correct all instances of "Adol-C" (proposed by jppeltere

Re: [deal.II] Design question about multiscale FEM with deal.ii

Many thanks, Wolfgang. Yes, I need to solve for the basis in each coarse cell separately, i.e., I have as many local fine meshes as coarse cells. I think that's actually an important point you are making here, maybe inadvertently: You have as many *meshes* as there are coarse cells. Now,

Re: [deal.II] Design question about multiscale FEM with deal.ii

On Wednesday, August 29, 2018 at 3:51:35 PM UTC+2, Jean-Paul Pelteret wrote: > > Hi Kendra, > > > Thank you, Jean-Paul, for the hint. Unfortunately this is not really > what I am looking for. I found a solution (that I don't like) by iterating > through two meshes on different levels simultaneou

Re: [deal.II] how to remove rows from system matrix

Thanks again this example will help me in future. Am I missing something or how FE_Enriched would help me to handle a singular matrix? I thought that this class just helps to integrate over discontinuous domains and it has nothing to do with zero degrees of freedom that I have in voids. среда,

Re: [deal.II] Design question about multiscale FEM with deal.ii

Hi Kendra, > Thank you, Jean-Paul, for the hint. Unfortunately this is not really what I > am looking for. I found a solution (that I don't like) by iterating through > two meshes on different levels simultaneously. Once I find a good solution I > will post it here and link to the code. I read

Re: [deal.II] Design question about multiscale FEM with deal.ii

On Wednesday, August 29, 2018 at 11:50:57 AM UTC+2, Jean-Paul Pelteret wrote: > > Hi Konrad, > > Do you think it would be possible to (ab)use the multigrid classes for > that? As far as I understand there are dofhandlers on each level of the > mesh. Can I get dofhandlers for all active subcells

Re: [deal.II] how to remove rows from system matrix

> As for examples of the FE_Enriched class, although there are no tutorials > that use it as of yet you can still look in the test directory for usable > code snippets, e.g. tests/fe/fe_enriched*.cc > I see that there is an equivalent of

Re: [deal.II] how to remove rows from system matrix

Hi, > I started with the code published in > https://journals.ub.uni-heidelberg.de/index.php/ans/article/view/22317 and > adopted it for my problem. Since you had removed Step-47 tutorial it is hard > to figure out how to use FE_Enriched class. FYI. It looks like step-47, which was removed in t

Re: [deal.II] how to remove rows from system matrix

> Well, I guess you could search for all support points of enriched DoFs > that lie outside of your domain of interest and add a homogeneous > constraint for them to the AffineConstraints object. The > DoFTools::map_dofs_to_support_points >

Re: [deal.II] Design question about multiscale FEM with deal.ii

Hi Konrad, > Do you think it would be possible to (ab)use the multigrid classes for that? > As far as I understand there are dofhandlers on each level of the mesh. Can I > get dofhandlers for all active subcells of say coarse cell number n in level > 5? The DoFHandler::begin_active()

Re: [deal.II] Design question about multiscale FEM with deal.ii

On Wednesday, August 29, 2018 at 4:23:32 AM UTC+2, Wolfgang Bangerth wrote: > > Since you really only want to interpolate between two meshes, I would just > have two DoFHandlers and two Triangulations and use the class InterGridMap > to > translate between the two of them. > > Or do you want t

[deal.II] Re: Inhomogeneous neumann bc for step-20

if you formulate a weak form for your problem you will see a boundary integral from grad(u), in step-20 it is equal to zero, since you changed the BC you should take into account this term in a weak form too. -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum opt

Re: [deal.II] how to remove rows from system matrix

Hi, > >What does the literature suggest how to handle this situation? > > The literature says, I cite: "The nodes that lie inside the void and whose > nodal support is > not intersected by the void are removed from the calculations. This is > performed usually by removing > the DOFs associ

Re: [deal.II] how to remove rows from system matrix

>What does the literature suggest how to handle this situation? > The literature says, I cite: "The nodes that lie inside the void and whose nodal support is not intersected by the void are removed from the calculations. This is performed usually by removing the DOFs associated with those