Warnings when setup and build a case

[dongxthu]

Hi E3SM team,

I can successfully run an air-land coupled case (compset: F20TR, res: ne30pg2_EC30to60E2r2). However, some strange warnings show when I setup and build the case (see attached file "setup & build.txt" for the whole screen output).

(1) when I setup, it shows

Create namelist for component mosart
Calling /scratch/general/vast/u6063131/E3SM/components/mosart//cime_config/buildnml
MOSART::buildnml compset 20TR_EAM%CMIP6_ELM%CNPRDCTCBCTOP_MPASSI%PRES_DOCN%DOM_MOSART_SGLC_SWAV_SIAC_SESP
False
MOSART::buildnml final infile_text is:
2025-01-24 20:02:44 glc

(2) when I build, it shows:

Component gptl build complete with 1 warnings
Component mct build complete with 42 warnings
Component csm_share build complete with 17 warnings

And in the log files, they show (attached):

gptl: Warning: Type mismatch between actual argument at (1) and actual argument at (2) (INTEGER(4)/LOGICAL(4)).
met: Warning: Type mismatch between actual argument at (1) and actual argument at (2) (CHARACTER(1)/INTEGER(4)).
Warning: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
...
cam_share: Warning: extra tokens at end of #include directive
Warning: Rank mismatch between actual argument at (1) and actual argument at (2) (rank-1 and scalar)
...

Note that in this case I use gcc/11.2.0. If I use gcc/8.5.0, these warnings when building will not show, but an error will show when building and make it fail to build (Attached file "csm_share.bldlog_gcc8.5.0.txt"):

warning: #warning "DSFMT_MEXP is not defined. I assume DSFMT_MEXP is 19937."
Error: Keyword argument ‘prev_limit’ at (1) is not in the procedure;

Although the case can run successfully and create reasonable output, I was wandering if these warnings affect the model results?

Thank you!!

setup & build.txt

gptl.bldlog.250124-201100.txt

mct.bldlog.250124-201100.txt

csm_share.bldlog.250124-201100.txt

csm_share.bldlog_gcc8.5.0.txt

[bartgol]

gptl: Warning: Type mismatch between actual argument at (1) and actual argument at (2) (INTEGER(4)/LOGICAL(4)).met: Warning: Type mismatch between actual argument at (1) and actual argument at (2) (CHARACTER(1)/INTEGER(4)). Warning: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1) ... cam_share: Warning: extra tokens at end of #include directive Warning: Rank mismatch between actual argument at (1) and actual argument at (2) (rank-1 and scalar) ...

I can't comment on the other ones, but these are common warnings. Perhaps a better investigation would show the issue, but as far as I can tell, these warnings (coming from MPI interfaces for F90) seem innocuous. It's as if the compiler picked the wrong procedure given the inputs. I think for gnu compilers e3sm should default to use the flag -fallow-argument-mismatch, which should silence that warning.

[dongxthu]

Hi Luca,

Thank you for your reply! I'm glad they are innocuous. Yes, I also saw this in other issues, for example, #4817. I tried to add following in the .bashrc, however, warnings still show. I was wondering if I made a mistake?

export FFLAGS="-I$INSTALLAREA/include -I$HDF5HOME/include -fPIC -O2 -fallow-argument-mismatch -fallow-invalid-boz"
export FCFLAGS="-I$INSTALLAREA/include -I$HDF5HOME/include -fPIC -O2 -fallow-argument-mismatch -fallow-invalid-boz"

Also, I was wondering if following output during setup is normal?

MOSART::buildnml compset 20TR_EAM%CMIP6_ELM%CNPRDCTCBCTOP_MPASSI%PRES_DOCN%DOM_MOSART_SGLC_SWAV_SIAC_SESP
False
MOSART::buildnml final infile_text is:

Thank you again for your help!

[rljacob]

The MOSART output is also harmless and normal.

[dongxthu]

Hi Robert, that's so great, thank you!! :)

[bartgol]

In E3SM, we do set CMAKE_<LANG>_FLAGS in our cmake files. These vars, if set, may overrule anything you have defined in your env (I am not entirely sure thoguh). I suppose you can check the build logs to se if the flag made it. If not, the best way is to set the flag in one of the cmake_macro/xyz files, corresponding to the mach/compiler you are using.

[dongxthu]

Hi Luca, I've found the file and now I understand what you mean. Thank you so much for your help!!