-
Notifications
You must be signed in to change notification settings - Fork 227
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
mom6/examples/coupled_AM2_LM3_SIS2/AM2_SIS2_MOM6i_1deg/INPUT is incomplete #49
Comments
It's working for me. Is this XML-driven? Any idea what file is missing. I'm guessing it might be this file but I'm not sure since it's out of our usual domain... |
When I list that INPUT I get the following. There seems to be a lot of files missing, not just mg_drag.res.nc, e.g., cns_*.nc What do you get?
|
The files are listed under https://github.com/CommerceGov/NOAA-GFDL-MOM6/tree/dev/master/examples/coupled_AM2_LM3_SIS2/AM2_SIS2_MOM6i_1deg/INPUT It's areal shame that FRE doesn't let you do a wild-card. FWIW here is my
|
This has nothing to do with FRE. Here's the procedure to reproduce:
|
@nikizadehgfdl you're correct; the INPUT/ directory is incomplete in the repository even though @Hallberg-NOAA and I both have complete directories. I think the cause can be traced back to the .gitignore created in the first commit of the repo which has a *.nc in it. We were still learning at that point. Ironically, commit e11a83f (from March) was explicitly trying to fix this setup but evidently I failed to test it properly. I'll fix the .gitignore first (which shouldn't be so inclusive) and then sort out the INPUT directory. |
- Issue #49 root cause can be traced to a global *.nc in the top-level .gitignore which hides inconsistency in the INPUT/ directory. - Added comments to .gitignore
…d_analysis_edits Layer ave T&S to diag_table; sea ice analysis mod
…s_in_nupoc Add two halo updates for taux and tauy in mom_surface_forcing_nuopc
- This is a simple fix for issue mom-ocean#49 - 1st issue The crash in prod+openmp mode comes from a non-thread-safe subroutine call from within an OMP PARALLEL DO region. The offending subroutine calls are advect_tracers_thicker() in ice_transport.F90. This subroutine has a pointer inside (Tr) that could change unexpectedly from underneath a thread depending on an input argument. The easy way to bypass the issue is to put the subroutine calls inside a OMP CRITICAL section (see below) which forces the threads to pass through one at a time. Of course this is not the efficient way to fix the issue. For an efficient fix we need to change the way the subroutine uses the pointer (Tr). - 2nd issue The crash with debug + openmp comes from another non-thread-safe block of code that allocates some arrays inside a OMP parallel region in SIS_slow_thermo.F90. The fix is to protect the block with the $OMP SINGLE clause to let only one thread to pass through
Renamed the framework module MOM_scaling_check.F90 to MOM_unique_scales.F90 to help differentiate it from MOM_check_scaling.F90, and renamed the subroutine check_scaling_factors() as check_scaling_uniqueness(). Also added _Dimensional_consistency.dox to describe the dimensional consistency testing. This commit should address the issues raised in the review of MOM6 PR mom-ocean#49. All answers and output are bitwise identical.
Renamed the framework module MOM_scaling_check.F90 to MOM_unique_scales.F90 to help differentiate it from MOM_check_scaling.F90, and renamed the subroutine check_scaling_factors() as check_scaling_uniqueness(). Also added _Dimensional_consistency.dox to describe the dimensional consistency testing. This commit should address the issues raised in the review of MOM6 PR mom-ocean#49. All answers and output are bitwise identical.
Remove logic to get and set CPU affinity for Darwin (MacOSX) platforms
It is missing many input files for ATM. Tests won't run:
FATAL from PE 26: mg_drag_init: source_of_sgsmtn="computed", but topography data file does not exist
The text was updated successfully, but these errors were encountered: