ffc team mailing list archive

["Garth N. Wells" <g.n.wells@xxxxxxxxxxxxxxx>]

Yep, I make a "dummy" element using ffc with the appropriate number of
components, then with this element I define a function, giving the function the
appropriate chunk of the solution vector. It would be nicer if a new function
could be 'extracted' from an existing function for given components.

Garth


Quoting Anders Logg <logg@xxxxxxxxx>:

> Great. We should probably add some functionality on the DOLFIN side to
> deal with subvectors/components of Functions so one can easily solve a
> mixed system and save the different parts/components as different
> fields for visualization. I assume you do this manually now?
> 
> /Anders
> 
> On Mon, Sep 26, 2005 at 05:09:34PM +0200, Garth N. Wells wrote:
> > That fixes the problem. 
> > 
> > Also, I've tested a reasonably complex mixed problem using the new fcc
> mixed
> > element functionality against a previous implementation using a vector, and
> the
> > results are identical.
> > 
> > Garth 
> > 
> > 
> > Quoting ffc@xxxxxxxxxx:
> > 
> > > Commit from logg (2005-09-26 17:02 CEST)
> > > ----------------
> > > 
> > >  - Fix bug in dofmap() for equal order mixed elements
> > >  - Add compiler option -d debuglevel
> > > 
> > > Garth, see if this helps.
> > > 
> > > Also, there is now a new command-line flag -d debuglevel. If > 0, more
> > > debug info will be printed during compilation (level 1 is reasonable).
> > > Also, ffc won't catch exceptions with this option.
> > > 
> > >   ffc  ChangeLog                            1.80
> > >   ffc  src/bin/ffc                          1.23
> > >   ffc  src/ffc/common/debug.py              1.4
> > >   ffc  src/ffc/compiler/dofmap.py           1.21
> > >   ffc  src/ffc/compiler/referencetensor.py  1.12
> > > 
> > > _______________________________________________
> > > FFC-dev mailing list
> > > FFC-dev@xxxxxxxxxx
> > > http://www.fenics.org/cgi-bin/mailman/listinfo/ffc-dev
> > > 
> > 
> 
> -- 
> Anders Logg
> Research Assistant Professor
> Toyota Technological Institute at Chicago
> http://www.tti-c.org/logg/
> 
> _______________________________________________
> FFC-dev mailing list
> FFC-dev@xxxxxxxxxx
> http://www.fenics.org/cgi-bin/mailman/listinfo/ffc-dev
>