| Thread Previous • Date Previous • Date Next • Thread Next |
On 06/10/2012 05:52 PM, Patrick Farrell wrote:
On 10/06/12 12:25, Simon Funke wrote:To make a long story short, I think we can only compute gradients with respect to dicretised objects - and that is not the case for Constant/Expression objects.Thinking about it some more, all objects in UFL are inherently discretised -- right? If I write log(x), that's a pointwise map from the function space X -> X. Even Expression classes implicitly have an associated function space, I think (Marie?).
All objects in UFL are not inherently discretized I think. But perhaps you did not mean _all objects_?An Expression in DOLFIN does not need to have a function space/finite element associated with it. However, if it is used in a form, its representation as a ufl.Coefficient, rather than as a dolfin.GenericFunction, is invoked, and its interpolant into some finite element function space will be used for the evaluation of the form. This space is deducted automatically/heuristically if no element is specified. However, the user can specify the element himself in which case this
element is used. -- Marie
| Thread Previous • Date Previous • Date Next • Thread Next |
