Add dependencies to CCPP metadata#153
Conversation
…-physics for code review and testing
…or testing the metadata-dependency implementation, fix an error for RRTMGP files
….py, remove legacy options that are no longer needed
… from ccpp/config/ccpp_prebuild_config.py
…rgument-mismatch for gfortran-10 compatibility
…amework and ccpp-physics
| dimensions = (horizontal_dimension) | ||
| type = real | ||
| kind = kind_phys | ||
| active = (flag_for_mass_flux_deep_convection_scheme == flag_for_gf_deep_convection_scheme) |
There was a problem hiding this comment.
I am curious how the active is used? Will the variable be allocated if active is false or this attribute will only be used by the ccpp/framework?
There was a problem hiding this comment.
The attribute is currently only used by the ccpp-framework to perform or skip certain operations in the auto-generated caps. In the future, data entirely internal to physics (i.e. everything in the GFS_interstitital DDT) will be allocated automatically by the ccpp-framework, and then active can be used to skip certain allocations.
But for the original/other GFS DDTs in GFS_typedefs.F90, it will only be used in the auto-generated caps. That's why the Fortran allocation logic in GFS_typedefs.F90 needs to be consistent with the active attribute in GFS_typedefs.meta for those data.
* update in fv_nudge to fix 666 label logic
This PR removes dependency information from the CCPP prebuild config. See NCAR/ccpp-framework#308 and NCAR/ccpp-framework#317 for details on the motivation for this change and the actual implementation. It also removes some legacy code used by the dynamic CCPP build in the past.
This PR also contains the changes in #156, i.e. the completion of adding the
activeattribute toGFS_typedefs.F90. On top of this PR, the missingactiveattribute forphy_fctdis added.Additionally, gfortran-10 compiler flags are added to CCPP's
CMakeLists.txt.Associated PRs:
earth-system-radiation/rte-rrtmgp#85
NCAR/ccpp-framework#317
NCAR/ccpp-physics#483
#153
ufs-community/ufs-weather-model#180
See ufs-community/ufs-weather-model#180 for regression testing.