Remove Immersed map from Adapting the ImmersedBoundaryGrid
#3690
Conversation
| @inline active_interior_map(grid::NamedTupleActiveCellsIBG, ::Val{:east}) = grid.interior_active_cells.east | ||
| @inline active_interior_map(grid::NamedTupleActiveCellsIBG, ::Val{:south}) = grid.interior_active_cells.south | ||
| @inline active_interior_map(grid::NamedTupleActiveCellsIBG, ::Val{:north}) = grid.interior_active_cells.north | ||
| @inline active_interior_map(grid::ActiveZColumnsIBG, ::Val{:surface}) = grid.active_z_columns |
There was a problem hiding this comment.
I don't understand the purpose of this system of functions or what a "named tuple active cells ibg" means, for example.
Also the function is called "map", but the object is called "cells". Which is it -- a map, or cells?
|
I think this is a good change. It might be nice to change some of the names of the functions as well, because I'm having a hard time understand the logic of them |
…Oceananigans.jl into ss/remove-immersed-map-from-adapt
…Oceananigans.jl into ss/remove-immersed-map-from-adapt
simone-silvestri
added
the
🚨 DO NOT MERGE 🚨
|
This PR should be ready, but I would like to add a test before merging it. I am planning to add it after next week, with more in-depth testing for the |
simone-silvestri
removed
the
🚨 DO NOT MERGE 🚨
|
I have added some tests, and everything is passing, including some tests for distributed immersed boundary grids that compare |
This PR removes the adapting of the immersed map from the immersed boundary grid by shuffling some definitions around and making sure that the active cells map is always passed as an argument instead of being embedded in the grid.
This can possibly help with parameter space issues in complex kernels (see ClimaOcean, Issue#116).
This PR is still a draft because I would like to take the opportunity to add some docstring for the immersed map