Developments shopping list

[esclapez]

• Refactor problem setup Add InitialBoundaryCondition (IBC) struct. #477
• Enable inflow on EBs #473
• Add on-the-fly diagnostics: PDF, conditional mean, iso-surface, ...
• Extend CI
• Extend V&V
• Add more tutorials
• Machinery to control dt when adding levels
• Add option to prune covered boxes when EB
• Time averaged data

[baperry2]

@esclapez What's the status of EB-inflow capability? I know you had a branch with this at least partially set up (https://github.com/AMReX-Combustion/PeleLMeX-archive/tree/EBInflow_Updated). Was that working or is more development needed? We have a case we'd like to run that would need EB-inflow capability.

[esclapez]

@baperry2 It's been on the back burner mostly because it is pretty invasive, but it is working. @EnnaDelfen is using it on one of her case (brought up to date with LMeX from a few months ago in her fork), even though she would like turbulence on EB-inflow ... always more !
The way it is implemented uses a new struct that handles all the EB-related needs (custom geom, EB state and so on). To avoid having to define everything when you have plain EB it uses SFINAE which I'm still on the fence about whether it is a good solution or not. I can try to refresh this, see if I can minimize the APIs breaks and open a draft PR so that we can start a discussion.
On the long run, I think it would be nice gather all the initial/boundary functions and data in a struct (maybe common to both C and LMeX ?)

[baperry2]

Thanks Lucas! Good to know this is working. I created a branch that rebases the EB-Inflow work to current development over here and verified that the test case still works as expected: https://github.com/baperry2/PeleLMeX/tree/ebinflow-rebase2

I think it would be good to get this merged in but having the up to date branch is enough for us to get started on the project that will use this capability. If you have a bit of time to make any modifications you'd like and create a draft PR, that would be awesome. Otherwise, I'll try to work toward getting this merged at some point.

I agree with the goal of eventually pushing all the problem-dependent BC/IC stuff into a single struct. We moved in that direction slightly in that direction for PeleC in AMReX-Combustion/PeleC#743, but there's still a lot to do there. I think the codes are different enough that they couldn't share this struct, but each code have their own.

[baperry2]

Just messing around a bit here with github's new sub-issue functionality. Not sure yet if it's useful.