ufs/dev PR295

[dustinswales]

SOURCE: Dustin Swales NOAA-GSL

DESCRIPTION OF CHANGES:
SCM changes identical to FV3UFS NOAA-EMC/fv3atm #950.

ISSUE:
N/A

ASSOCIATED PRs:

• NCAR/ccpp-physics #1149

[scrasmussen]

Approving because you said there are fewer changes here but pointing them out just to double check, in FV3UFS NOAA-EMC/fv3atm #950 we have the following class of changes that aren't in this PR

• get_tracer_index -> get_physics_tracer_index
• integer, parameter :: physics_no_tracer = -99

[grantfirl]

Approving because you said there are fewer changes here but pointing them out just to double check, in FV3UFS NOAA-EMC/fv3atm #950 we have the following class of changes that aren't in this PR

• get_tracer_index -> get_physics_tracer_index
• integer, parameter :: physics_no_tracer = -99

Yes, I think that everything is good. FV3 and SCM now use different solutions for the get_tracer_index thing. The SCM has it defined in GFS_typedefs, but the FV3 uses the one from FMS (which the SCM doesn't use).

[grantfirl]

The debug mode failure should be debugged at some point, but not here. It doesn't fail on any other platforms.

[scrasmussen]

The debug mode failure should be debugged at some point, but not here. It doesn't fail on any other platforms.

Agreed, I was able to reproduce this error locally which will make it easier to debug. I'll track it down.

[grantfirl]

The debug mode failure should be debugged at some point, but not here. It doesn't fail on any other platforms.

Agreed, I was able to reproduce this error locally which will make it easier to debug. I'll track it down.

Oh, good. I was not able to reproduce a failure locally. Which version of GNU were you able to reproduce with? I have 12.4 on my Mac.

[scrasmussen]

The debug mode failure should be debugged at some point, but not here. It doesn't fail on any other platforms.

Agreed, I was able to reproduce this error locally which will make it easier to debug. I'll track it down.

Oh, good. I was not able to reproduce a failure locally. Which version of GNU were you able to reproduce with? I have 12.4 on my Mac.

I'm using 12.2 on Derecho (by locally I meant off of Github lol).

Using a debugger I found the specific place where the code is breaking due to NaNs, still working on why though. PR #607 has that branch.