Ghost cell allocation for AMR and 3rd order reconstruction

[kyriakosTapinou]

Will there be an unintended error when specifying more ghost cells than is required for the spatial order of a simulation?

For example, if I set the spatial reconstruction to 3rd order, but want to use AMR I would need 4 ghost cells set.

[tomidakn]

Mesh refinement requires even number of ghost cells (you should get an error message if you set nghost=3 with mesh refinement). So you do need 4 ghost cells for 3rd order reconstruction.

But please also note that currently the refinement operation is only 2nd order.

[kyriakosTapinou]

@tomidakn yes the error pops up when there is a mismatch.

I was wondering if there was anything behind the scenes which might be an issue - from what I could see relating to block bounds etc there should not be an issue because of how the ks and ke start from the block domain edge but I wanted to check if there was something obvious I was missing.

Ok, I did not know the order was only 2nd for refinement.

Thank you!

[tomidakn]

I do not think there is any issue. Please let me know if you have any specific issue/concern.