search

NCAR / ccpp-physics

GFS physics for CCPP
Other
58 stars 145 forks source link

Fix RAP (Grell Freitas) decomp b4b issues #942

Closed SamuelTrahanNOAA closed 2 years ago

SamuelTrahanNOAA commented 2 years ago

This PR replaces https://github.com/NCAR/ccpp-physics/pull/933 and is identical, except that it comes from another fork. All of this work comes from @climbfuji

Bug fixes in physics/cu_gf_deep.F90 and physics/cu_gf_driver.F90 for bit-for-bit identical results with GF when changing the MPI decomposition. Instead of using the pefc value of last column (pefc was a scalar), which is incorrect, turn pefc into a vector of length horizontal_loop_extent and use the correct value in GF deep convection.

Minor formatting cleanup in removal of unused symbols/code.

Associated PRs:

https://github.com/NOAA-EMC/fv3atm/pull/553 https://github.com/ufs-community/ufs-weather-model/pull/1257

For regression testing, see https://github.com/ufs-community/ufs-weather-model/pull/1243

SamuelTrahanNOAA commented 2 years ago

I didn't do it. Dom did. I'm replacing his PR so he can use his 4 hours a week for something more critical.

climbfuji commented 2 years ago

Thanks for taking care of getting those changes in @SamuelTrahanNOAA!