On 9/11/19 11:03 AM, Konrad wrote:
>
> That indeed depends on the geometry of the domain and it is not an
> obvious result. Example: Think of a 3D hyper-shell with an inner and
> outer radius. There is a one-dimensional space of non-trivial radial
> vector fields (with respect to the origin) th
Hi Wolfgang,
> > Using approach 1.) I get timing results like that:
> > (using MPI with Petsc, AMG preconditioning for *A* on one machine with
> > four cores, no preconditioning for the Schur complement *S =
> > -B*A^{-1}*B^T + C* from 1.)
> >
> > Running using PETSc.
> > Number of active
On 9/11/19 7:04 AM, Konrad wrote:
>
> The system that needs to be solved formally reads
> *(P) A sigma + B^T u = 0
> *
> * B sigma + C u = f*
>
> *A* is sort of a mass matrix so it is positive (corresponds to
> ** in *H(curl)*) and *C* is positive semi-definite
> (corresponds to ** in
Dear deal.ii community,
Apologies for so many questions lately... but here is another one. This one
is a bit technical and aims at good solvers. In a simplified form you can
formulate it this way:
I implemented a PDE solver with a for the 3D vector Laplacian *curl(curl u)
- grad(div u) = f*. I
