Closed brhillman closed 4 years ago
@whannah1 that's a good point. But in this case, I don't think I want to wait for the changes to be accepted into E3SM and trickle down if we have any hope of using this for INCITE (we might not anyways). I can open a PR with the identical changes in E3SM though, because I believe this is a problem there too.
@brhillman, ok, I agree with this approach.
Fix runtime failures when model is build with PGI or GNU. To do this, we need to only set nswbands and nlwbands in radconstants (as parameters). Previously, we had tried to set these dynamically in radiation by reading the values from the kdist objects, which are populate from data read from netcdf files at initialization, but some routines seemed to need these dimension sizes to be compile-time constants. So, unfortunately, we revert to making the values of nswbands and nlwbands hard-coded. Fixes #110