Update CAM-SIMA to run Kessler and Held-Suarez bit-for-bit with CAM #232
Conversation
nusbaume
added
enhancement
|
The test failures in python 3.12 are due to issues with ParamGen in CIME using a deprecated package. I'll try to get that fixed and the CIME external updated, but if I don't get there in time I'll just remove python 3.12 from the testing options, at least for now. So feel free to ignore those failures when reviewing. Thanks! |
|
All test failures have been fixed now, so this PR should be ready for review. Thanks! |
| @@ -260,6 +265,9 @@ subroutine air_composition_init() | |||
| thermodynamic_active_species_kc(icnst) = kc1 | |||
| icnst = icnst + 1 | |||
| dry_species_num = dry_species_num + 1 | |||
| !Notify constituent object that this species is | |||
| !thermodynamically active | |||
| call const_set_thermo_active(idx, .true.) | |||
There was a problem hiding this comment.
Just curious, what is the default if this were not specified?
There was a problem hiding this comment.
The default is .false., which is specified in the constituents object itself:
| ! pref_mid_norm | ||
| call mark_as_initialized("reference_pressure_normalized_by_surface_pressure") | ||
| call mark_as_initialized("reference_pressure_in_atmosphere_layer_normalized_by_reference_pressure") |
There was a problem hiding this comment.
I was questioning in in this standard name. I see that this variable does not currently exist. Should it be of instead so it matches the non-normalized name? If so, you'll need to change it everywhere it is used.
There was a problem hiding this comment.
You are definitely correct that there is a standard name bug here, but the error isn't with this standard name, but the one before it, which should be reference_pressure_in_atmosphere_layer instead of reference_pressure_of_atmosphere_layer.
This is because the in_atmosphere_layer version is already in the standard names dictionary and has been accepted by the scientists according to our spreadsheet. I've thus gone ahead and replaced reference_pressure_of_atmosphere_layer with reference_pressure_in_atmosphere_layer in CAM-SIMA.
There was a problem hiding this comment.
Just looking for some clarification on a few things but otherwise, looks great @nusbaume!!
There was a problem hiding this comment.
thanks Jesse!
I also resolved @cacraigucar 's comments that you addressed (but left the comments that Cheryl may want to see the answer to when she returns).
nusbaume
dismissed
cacraigucar’s stale review
@cacraigucar's requested changes have been applied, and reviewed by a different SE. Plus this reviewer is currently on PTO and we don't want to wait until they are back.
|
@mwaxmonsky if you get the chance can you approve or re-review this PR? I am trying to avoid breaking too many of my own rules. |
|
@mwaxmonsky it looks like you resolved your concerns earlier, and since I am hoping to have this repo ready for the presentation in the next 40 minutes I am going to go ahead and break my own rules and get this merged in. Of course feel free to still review the code if you want and let me know if there is anything problematic (as we can always open another PR). Thanks! |
This PR introduces a large number of standard name and other changes to allow one to run the new Kessler and Held-Suarez physics suites with CAM snapshots, and to ensure that they are bit-for-bit with CAM's results.
This PR also ensures that the SE dycore still runs with both Held-Suarez and Kessler physics suites (but it's values can't be formally checked yet due to the dycore being out-of-sync with CAM).
This PR also performs some code clean-up, and adds the ability for
check_fieldto print the location (as an MPI rank, array column, and vertical level) of the maximum difference for each variable, which should help with debugging.closes #210
Tests run
Ran single timestep
FPHYStestruns with CAM snapshots for Held-Suarez and Kessler with GNU on Derecho with the ne3pg3 grid, and got bit-for-bit results. Also ranFHS94andFKESSLERcompsets with thene5pg2grid and the GNU compiler for five timesteps each, and both ran successfully.