Hello Daniel, Thanks for the prompt reply. Yes, I thought about this as deal.ii uses P4est which does not support anisotropic refinement.
I wanted to confirm that this was indeed the case and I am not missing something. Thanks again for the clarification. On Sunday, March 5, 2017 at 7:48:56 AM UTC-5, Daniel Arndt wrote: > > Rajat, > > > I am using the GridGenerator::hyper_rectangle function to create a >> rectangular grid with 1000 X 500 X 1 (0.5 million) elements. >> >> My question is: - since the coarsest mesh contains 0.5 million elements, >> does that mean all the processors have 0.5 million elements? >> (I think the answer is yes, I read that the deal.ii (p4est) stores the >> coarsest mesh on all processors.) >> > Yes, that is true. > > I need a mesh (overall) which should be rectangular and should have 1000 X >> 500 X 1 elements. Note: It should only have 1 element in the z-direction. >> Is there any way such that I can have this mesh by not replicating the >> whole mesh on all processors as it takes a lot of memory. I thought of >> making a 2D mesh and then refining it and then extruding it using >> extrude::triangulation, but extrude triangulation does not work on refined >> meshes. >> > As you already said, we (currently) require to store the same coarse mesh > on all processes and only isotropic refinement is allowed. This means that > you can get 1 element in z-direction if you don't refine. There is some > work [1] on a Triangulation that allows you to specify the distribution of > cells across processes yourself without the need of a common coarse mesh. > > Best, > Daniel > > [1] https://github.com/dealii/dealii/pull/3956 > -- 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 [email protected]. For more options, visit https://groups.google.com/d/optout.
