Roy ninja'ed me here, but I did want to add a couple of things: On Mon, Mar 7, 2016 at 4:26 PM, Roy Stogner <royst...@ices.utexas.edu> wrote:
> > On Mon, 7 Mar 2016, Giorgio Bornia wrote: > > > in the list of supported elements in libmesh there is no Tri7. > > Then, I guess it's not possible to implement things such as > > Crouzeix-Raviart pairs, right? > > Or maybe it is in a DOF sense without underlying Tri7 topology, > > then printing back to Tri6 or Tri3. > > In the DOF sense, we handle bubble shape functions on triangles by > associating them with the element rather than with any of its nodes. > See, for example, the hierarchic finite elements. > (Our real Achilles' heels are tets and pyramids, where we currently > have no options for dofs on triangular sides) > To clarify, this is not a fundamental limitation of the library, per se, but rather a limitation of manpower/need to implement those geometric element types to support the finite element types. If you're wanting to add these, we'd love it! I'd be happy to guide you. It's not terribly difficult, but it does require interacting with the guts of the library. > You're right that we have to decimate before we output any > sophisticated elements in a non-native format. > See for example how FENedelecOne is handled for the Nedelec elements of the first type. Best, Paul ------------------------------------------------------------------------------ Transform Data into Opportunity. Accelerate data analysis in your applications with Intel Data Analytics Acceleration Library. Click to learn more. http://makebettercode.com/inteldaal-eval _______________________________________________ Libmesh-users mailing list Libmesh-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/libmesh-users
