Jobcomp Error: Rank mismatch

compilation-installation
1

Hello everybody! I hope they are fine!

I have a problem compiling my simulation. I’m just running a simulation of physical parameters for the Gulf of California inlet area. Previously it ran without any problem, but lately, I get the following error:

% ===================================================================================================
/usr/local/netcdf-f/include/netcdf.inc:1277:38: warning: missing terminating ’ character [-Winvalid-pp-token]
integer nf_ecantread ! Can’t read.
^
/usr/local/netcdf-f/include/netcdf.inc:1279:38: warning: missing terminating ’ character [-Winvalid-pp-token]
integer nf_ecantwrite ! Can’t write.
^
/usr/local/netcdf-f/include/netcdf.inc:1281:38: warning: missing terminating ’ character [-Winvalid-pp-token]
integer nf_ecantcreate ! Can’t create.
^
3 warnings generated.
gfortran -c -O3 -fdefault-real-8 -fdefault-double-8 put_global_atts_.f -o put_global_atts.o
put_global_atts_.f:1235:64:

1218 | & 1, theta_s)
| 2
…
1235 | & sc_w)
| 1
Error: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
put_global_atts_.f:1239:64:

1218 | & 1, theta_s)
| 2
…
1239 | & Cs_w)
| 1
Error: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
put_global_atts_.f:1242:62:

1218 | & 1, theta_s)
| 2
…
1242 | ierr=nf_put_att_double(ncid,nf_global,‘sc_r’,nf_ftype,N,sc_r)
| 1
Error: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
put_global_atts_.f:1245:62:

1218 | & 1, theta_s)
| 2
…
1245 | ierr=nf_put_att_double(ncid,nf_global,‘Cs_r’,nf_ftype,N,Cs_r)
| 1
Error: Rank mismatch between actual argument at (1) and actual argument at (2) (scalar and rank-1)
gmake: *** [Makedefs:14: put_global_atts.o] Error 1

% ===================================================================================================
The parameters I have been using are:

% Vertical grid parameters
%
theta_s = 7;
theta_b = 0.3;
hc = 10.;
vtransform = 2.;

I don’t know why right now I have this error. Someone can help me o guide me with my problem!?

2

Hi:

I am having this problem, too, with a fresh download of v1.1. It seems to be a new feature of the latest gfortran:

https://gcc.gnu.org/gcc-10/porting_to.html

You can add this option to the compiler options on line 145 of jobcomp:

-fallow-argument-mismatch

and be careful to preserve the ending whitespace there.

-Ed

3

Hi,
Thank you so much for your help Ed. I followed your recommendation, but I have the same error. Maybe I didn’t add it in the right place. These are my lines in jobcomp file:

compiler options

FC=gfortran
-fallow-argument-mismatch

Can you check if is it correct? please!
Thanks again!

4

I can’t tell from your message. It looks as if the “FC=” statement is inside a comment block.

Can you post your entire file as an attachment here? Alternately, you can email it to me directly at edwarddzaron@oregonstate.edu, and I will compare it with mine.

5

Thanks Ed, if you don’t mind I’ll email it to you because I don’t know how to attach it here!

6

Seems it is needed here (now compiles serial and openmo)

    elif [[ $FC == gfortran ]] ; then
            CPP1="cpp -traditional -DLinux"
            CFT1=gfortran
            FFLAGS1="-O3 -fdefault-real-8 -fdefault-double-8 -fallow-argument-mismatch"    
                                                             **************************
7

Seems it is needed here (now compiles serial and openmp)

    elif [[ $FC == gfortran ]] ; then
            CPP1="cpp -traditional -DLinux"
            CFT1=gfortran
            FFLAGS1="-O3 -fdefault-real-8 -fdefault-double-8 -fallow-argument-mismatch"    
                                                             **************************
8

And here for MPI (mpif90)

    FFLAGS1="$FFLAGS1 $MPIINC -fallow-argument-mismatch"