dolfin team mailing list archive

[Kristian Oelgaard <k.b.oelgaard@xxxxxxxxxx>]

Quoting Anders Logg <logg@xxxxxxxxx>:

> On Wed, Jul 15, 2009 at 11:49:32AM +0200, Kristian Oelgaard wrote:
> > Quoting phil marinier <lonewolf_13p@xxxxxxxxxxx>:
> >
> > >
> > > If I use   Function.vector().get()   to get my nodal values for a
> solution I
> > > obviously get a 1-dimensional vector. I made sure I used    .order()
> on
> > > my mesh so that the numbering would be UFC standard. What I can't seem to
> > > find is what the UFC global numbering standard is. I read parts of the
> UFC
> > > user manual, but all I could find was local entity numbering with respect
> to
> > > the global numbering.
> > >
> > > What I would like to end up with is a 2-dimensional array with all the
> nodal
> > > values. I can easily convert my vector into such an array if someone can
> > > point me to the numbering scheme. However, if there is an easier way to
> do
> > > this, please let me know.
>
> UFC does not dictate a global numbering for the degrees of freedom,
> only for the local numbering of mesh entities.
>
> If you explain which element you are using (like piecewise linears),
> someone can answer exactly how the degrees of freedom are numbered.

Or, you can just look at the tabulate_dofs() function in the generated code.
But if you change the element in the *.ufl file you'll have to change your
implementation accordingly, which is why I would suggest using one of the two
methods I proposed in an earlier post.

Kristian

> --
> Anders
>
>
> > I'm not sure what your end goal is, but if you want to manipulate the
> solution
> > vector one cell at a time I see two ways of doing it:
> >
> > Loop the cells
> >
> > 1) tabulate the dofs on each cell and the use the vectors get() methods
> with
> > those dofs as extra arguments.
> >
> > or
> >
> > 2) use the functions interpolate() method to interpolate the function
> values on
> > the current cell.
> >
> > I don't know which method is best/fastest, I use 1).
> >
> > Take a look at fem/Assembler.cpp.
> >
> > Kristian
> >
> > > Thanks for your help.
> > >
> > > Phil
> > >
> >
> >
> > _______________________________________________
> > DOLFIN-dev mailing list
> > DOLFIN-dev@xxxxxxxxxx
> > http://www.fenics.org/mailman/listinfo/dolfin-dev
>