Forgot to mention that I'm using LibMesh version 9.2.1. Are Heirarchic
basis functions fully implemented in that version?


> I'm afraid the ASCII art in include/geom/cell_hex27.h and the
> side_nodes_map and edge_nodes_map arrays in src/geom/cell_hex27.C is
> the best currently available documentation.


Thanks, I'll go through this.

OK, if you can share your code, someone with a SLEPc-enabled libmesh
> build handy might be able to confirm your results too.


I have SLEPc-enabled libmesh, I will check if I get the same results, and
also get the output from -ksp_monitor. The code is part of a bigger
software (NEMO5) and it's quite difficult to make a standalone.

On Thu, Jan 8, 2015 at 7:02 PM, John Peterson <[email protected]> wrote:

> On Thu, Jan 8, 2015 at 4:57 PM, Harshad Sahasrabudhe
> <[email protected]> wrote:
> >> It's also possible there's a bug with 4th-order hierarchics in 3D, the
> >> cube_indices() function in fe_hierarchic_shape_3D.C scares me.
> >
> >
> > I tried, but I don't understand what's happening inside the
> cube_indices()
> > function.
> >
> > 51  // The only way to make any sense of this
> > 52  // is to look at the mgflo/mg2/mgf documentation
> >
> > Where can I find this documentation for the cube_indices() function?
>
> I'm afraid the ASCII art in include/geom/cell_hex27.h and the
> side_nodes_map and edge_nodes_map arrays in src/geom/cell_hex27.C is
> the best currently available documentation.
>
>
> >> Are you sure it converged?  I know as you increase p, the condition
> >> number gets higher/problem gets harder to solve.
> >> Run your code with -ksp_monitor, etc.
> >
> > I'm using MKL FEAST for calculating the eigenvalues. It doesn't use an
> > iterative method and first calculates the number of eigenvalues in an
> > interval. I get a 0 at that step for an order 4 Hierarchic basis in HEX27
> > geometry.
>
> OK, if you can share your code, someone with a SLEPc-enabled libmesh
> build handy might be able to confirm your results too.
>
> --
> John
>
------------------------------------------------------------------------------
Dive into the World of Parallel Programming! The Go Parallel Website,
sponsored by Intel and developed in partnership with Slashdot Media, is your
hub for all things parallel software development, from weekly thought
leadership blogs to news, videos, case studies, tutorials and more. Take a
look and join the conversation now. http://goparallel.sourceforge.net
_______________________________________________
Libmesh-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libmesh-users

Reply via email to