[WIP] Add new flexible FV interpolation machinery. #31974
Conversation
grmnptr
force-pushed
the
fv-interpolation-31965
branch
from
28b49f7 to
3211d05
Compare
|
|
||
| /** | ||
| * Base class for face interpolation functions used by linear FV objects. | ||
| * Maybe we can extend it a little late to regular FV as well. |
There was a problem hiding this comment.
| * Maybe we can extend it a little late to regular FV as well. | |
| * Maybe we can extend it a little later to regular FV as well. |
| template <typename Derived> | ||
| static Real callInterpolate(const FVInterpolationMethod & method, |
There was a problem hiding this comment.
instead of CRTP you could use type erasure and a registry like @NamjaeChoi transitioned to for Kokkos. But I'm also ok with the CRTP
There was a problem hiding this comment.
I'd be curious about the details! Could you point me to the relevant parts?
There was a problem hiding this comment.
It's discussed (not super thoroughly) in https://mooseframework.inl.gov/syntax/Kokkos/. Let me know if you want more detail
| * the derived classes can use their own member variables (like a power variable for inverse | ||
| * distance weighting, or a limiter variable etc). | ||
| */ | ||
| class FVInterpolationMethod : public GeneralUserObject |
There was a problem hiding this comment.
I don't know, I just wanted something quick to start with. If I get the green light I will make it a regular mooseobject.
There was a problem hiding this comment.
When I hack I hack with user objects.
There was a problem hiding this comment.
You have my green light. Note that my light will always be greener if not using a user object haha. Then I know there's going to be first class support in the framework
|
I'm not happy with this design yet, will come back a little more. |
2 participants
Closes #31965