-
The Davies et al benchmark test looks to be a good test to see whether the solvers are working correctly. There is a table in the manual (Table 7) with values that can be compared to your model run, t…
-
In the `OperatorBase::vmult()` (or `apply()`) function, we treat constrained dofs as part of the linear system of equations with values of 1 on the diagonal of the "matrix". I do not see where we impo…
-
`VectorTools::project` has not been implemented for hp-adaptive applications in parallel yet. The problem surfaces when using a `QCollection` object for the project function.
Documentation says tha…
-
The newly introduced `CellWeights` class (#7305) provides a cell weighting functionality for both `p::s::Triangulation` and `p::d::Triangulation` implementations.
In case a `p::s::Triangulation` is…
-
https://github.com/dealii/dealii/pull/7744 reworked read/write interface to vectors from FEEvaluation and added several FIXMEs that we should fix. Those have to do with parallel PETSc/Trilinos that do…
-
I have been encountering failures to converge when using SUPG in 3D. We actually haven't been successful in running a single 3D model to completion with SUPG. This includes both 3D spherical geometry …
-
When I do not set boundary compositions for a specific model, the model runs fine, but after a while erroneous compositions flow in. When I set the boundary compositions to intitial compositions the f…
-
@tjhei This is one for you, pointed out to me by @marcfehling .
In `stokes_matrix_free.cc`, we apply the block preconditioner in the following way (abridged):
```
void
BlockSchurGMGPreco…
-
I have run into an issue with the GMG and DC Picard (and by extension the Newton solver). In short, `single Advection, iterated Stokes` with gmg works, `single Advection, iterated defect correction St…
-
While working on #16856 (especially on @peterrum's comment https://github.com/dealii/dealii/pull/16856#discussion_r1569096582) I noticed, that restriction function `MGTwoLevelTransfer::restrict_and_ad…