bshanahan
bshanahan
A minimum field line length is probably the best solution, but one could also check that a few extra guard cells wouldn't catch the field line.
This looks great! The only issue I can see at the moment is with something like DDX(FieldPerp), which would likely need 3D metrics if available.
Might I also suggest this would be a good time to implement nonuniform FFTs? If we implement support for different `nz`, then we might as well let that `nz` vary....
> Maybe simpler to just add to the error message that it might be due to this? Unfortunately this is not just an issue with this particular operator. It could...
I guess option 1 is the best, but this is more a question for @ZedThree
As most of the current users of Draco are using FCI, it's probably best to add openmp to the available libraries.
Here are some lines early on in the simulation: ``` 2.500e+04 1289 8.91e-01 67.3 7.1 4.4 3.8 17.4 2.510e+04 1130 7.86e-01 66.9 7.0 4.5 4.3 17.4 2.520e+04 1105 7.72e-01 66.8...
Integrated tests pass (when switching laplacian inversion to `PETSC` instead of `cyclic` -- a current limitation of 3D metrics) and unit tests pass except for `SNBConductionTest.CreateComponent` and `SNBConductionTest.Transform` -- both...
Looks like a bug to me.