dolfin team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

I think we should do arbitrary order isoparametric. We also need more
general support for non-affine mappings like Piola for H(div) or
H(curl) conforming elements.

/Anders


On Fri, Dec 16, 2005 at 05:20:00PM +0100, Garth N. Wells wrote:
> On Fri, 2005-12-16 at 10:12 -0600, Anders Logg wrote:
> > Yes, it would require some major work in quite a few places, FIAT,
> > FFC, DOLFIN, and it's not on my TODO list for the next couple of
> > months at least.
> > 
> 
> Is the plan to support arbitrary order isoparametric mappings, or just a
> limited number (linear and quadratic only, for example)? 
> 
> Garth 
> 
> 
> > /Anders
> > 
> > 
> > On Fri, Dec 16, 2005 at 05:05:23PM +0100, Luca Antiga wrote:
> > > Hi guys, thanks for the prompt reply.
> > > In fact, I was referring to curved elements (sorry for not being 100% clear),
> > > but nevertheless
> > > I found the Poisson example very informative. A really nice way to go.
> > > So, if I understand correctly, implementing curved tets would be on the same
> > > level as implementing
> > > bricks. Am I right?
> > > 
> > > Luca
> > > 
> > > On 12/16/05, Anders Logg <logg@xxxxxxxxx> wrote:
> > > 
> > >     Right.
> > > 
> > >     /Anders
> > > 
> > >     On Fri, Dec 16, 2005 at 09:49:25AM -0600, Robert C Kirby wrote:
> > > 
> > >     > No -- I think higher order geometry is what is meant - curved elements.
> > >     >  We don't do those.
> > >     >
> > >     > On Dec 16, 2005, at 9:47 AM, Anders Logg wrote:
> > >     >
> > >     > >As Garth points out, we don't support prisms (yet).
> > >     > >
> > >     > >For higher order elements, take a look at the example in
> > >     > >
> > >     > >    src/demo/fem/convergence/
> > >     > >
> > >     > >which solves Poisson's equation in 2D and 3D with order ranging
> > >     > >from q = 1 to q = 5.
> > >     > >
> > >     > >The only thing you need to change to do higher order elements is a
> > >     > >number
> > >     > >(and then recompile with FFC).
> > >     > >
> > >     > >/Anders
> > >     > >
> > >     > >
> > >     > >On Fri, Dec 16, 2005 at 10:45:07AM +0100, Luca Antiga wrote:
> > >     > >>Dear Dolfin developers,
> > >     > >>  I've looked into Dolfin today and it's a really impressive piece of
> > >     > >>software.
> > >     > >>Congratulations for your work.
> > >     > >>Just a curiosity: how hard it would be to code different element
> > >     > >>types (e.g.
> > >     > >>quadratic tetrahedra or prisms) into Dolfin?
> > >     > >>Thanks a lot
> > >     > >>
> > >     > >>Luca
> > >     > >>
> > >     > >>
> > >     > >>
> > >     > >
> > >     > >_______________________________________________
> > >     > >DOLFIN-dev mailing list
> > >     > > DOLFIN-dev@xxxxxxxxxx
> > >     > >http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
> > >     >
> > >     >
> > >     > _______________________________________________
> > >     > DOLFIN-dev mailing list
> > >     > DOLFIN-dev@xxxxxxxxxx
> > >     > http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
> > >     >
> > > 
> > > 
> > >     _______________________________________________
> > >     DOLFIN-dev mailing list
> > >     DOLFIN-dev@xxxxxxxxxx
> > >     http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
> > > 
> > > 
> > 
> > > _______________________________________________
> > > DOLFIN-dev mailing list
> > > DOLFIN-dev@xxxxxxxxxx
> > > http://www.fenics.org/cgi-bin/mailman/listinfo/dolfin-dev
> > 
> > 

-- 
Anders Logg
Research Assistant Professor
Toyota Technological Institute at Chicago
http://www.tti-c.org/logg/