Issue Running ROMS umaine cosine model

[shivaligangwar]

while running oceanM i am getting following error -
mpirun -np 96 /gpfs3/home/ibm/GOYAL/shivali/RKU/ROMS_kate_umaine/oceanM /gpfs3/home/ibm/GOYAL/shivali/RKU/ROMS_kate_umaine/indian_ocean.in

Model Input Parameters: ROMS/TOMS version 3.6
Tuesday - December 9, 2014 - 4:19:02 PM
-----------------------------------------------------------------------------
forrtl: severe (66): output statement overflows record, unit -5, file Internal Formatted Write
Image PC Routine Line Source
oceanM 000000000082F60B Unknown Unknown Unknown
oceanM 00000000007C91EE Unknown Unknown Unknown
oceanM 000000000042A486 Unknown Unknown Unknown
oceanM 000000000040EBD4 Unknown Unknown Unknown
oceanM 000000000040B778 Unknown Unknown Unknown
oceanM 000000000040B18B Unknown Unknown Unknown
oceanM 000000000040B036 Unknown Unknown Unknown
libc.so.6 0000003E4341ECDD Unknown Unknown Unknown
oceanM 000000000040AF29 Unknown Unknown Unknown

[kate]

I can't tell from this if it's related to the Cosine model or not. Can you recompile with USE_DEBUG on and try this again? It should give you line numbers and routine names (I hope).

[shivaligangwar]

Following is the error after enabling debug -

iitmlogin3:/gpfs3/home/ibm/GOYAL/shivali/RKU/ROMS_kate_umaine $ cat roms.error.100790
forrtl: severe (66): output statement overflows record, unit -5, file Internal Formatted Write
Image PC Routine Line Source
oceanG 00000000020B9FDB Unknown Unknown Unknown
oceanG 0000000001FAD846 mod_ncparam_mp_in 1643 mod_ncparam.f90
oceanG 0000000000526751 read_phypar_ 179 read_phypar.f90
oceanG 0000000000428B3B inp_par_ 90 inp_par.f90
oceanG 000000000040B956 ocean_control_mod 84 ocean_control.f90
oceanG 000000000040B202 MAIN__ 97 master.f90
oceanG 000000000040AF36 Unknown Unknown Unknown
libc.so.6 00002B81636F9CDD Unknown Unknown Unknown
oceanG 000000000040AE29 Unknown Unknown Unknown
iitmlogin3:/gpfs3/home/ibm/GOYAL/shivali/RKU/ROMS_kate_umaine $

[shivaligangwar]

Dear Sir,

Error seems to be in this loop -

!-----------------------------------------------------------------------
! Set passive tracers surface flux metadata. The variable name is the
! same as the basic tracer but with the _sflux suffix.
!-----------------------------------------------------------------------
!
DO i=NAT+1,MT
varid=varid+1
WRITE (6,*) 'Inside Netcdf subroutine varid MV -',varid,MV
IF (varid.gt.MV) THEN
WRITE (stdout,60) MV, varid
STOP
END IF
idTsur(i)=varid
DO ng=1,Ngrids
Fscale(varid,ng)=1.0_r8
Iinfo(1,varid,ng)=r2dvar
END DO
WRITE (6,*) 'Vname -', Vname(1,varid)
line :1643 ==>here WRITE (Vname(1,varid),'(a,a)') &
& TRIM(ADJUSTL(Vname(1,idTvar(i)))), '_sflux'
WRITE (Vname(2,varid),'(a,a)') &
& TRIM(ADJUSTL(Vname(2,idTvar(i)))), ', surface flux'
WRITE (Vname(3,varid),'(a,1x,a)') &
& TRIM(ADJUSTL(Vname(3,idTvar(i)))), 'meter second-1'
WRITE (Vname(4,varid),'(a,a)') &
& TRIM(Vname(1,varid)), ', scalar, series'
WRITE (Vname(5,varid),'(a)') &
& TRIM(ADJUSTL(Vname(5,idTvar(i))))
END DO

[kate]

OK, very good. So you added a write statement:

Code: Select all

WRITE (6,*) 'Vname -', Vname(1,varid)

What you really want is:

Code: Select all

WRITE (6,*) 'Vname -', i, Vname(1,idTvar(i))

What does that give you?

Something to check is your varinfo.dat file. Do you have all the cosine variables in there? I have a chunk starting with:

Code: Select all

!--------------UMAINE variables

[shivaligangwar]

Dear Sir,

I tried folllowing WRITE as suggested by you -
WRITE (6,*) 'Vname -', i, Vname(1,idTvar(i))

and output is -
Model Input Parameters: ROMS/TOMS version 3.6
Thursday - December 11, 2014 - 3:05:26 PM
-----------------------------------------------------------------------------
Inside Netcdf subroutine varid MV - 175 950
Vname - 3
NO3

Nothing there after and error log -

iitmlogin3:/gpfs3/home/ibm/GOYAL/shivali/RKU/ROMS_kate_umaine $ cat roms.error.101610
forrtl: severe (66): output statement overflows record, unit -5, file Internal Formatted Write
Image PC Routine Line Source
oceanM 00000000020BBADB Unknown Unknown Unknown
oceanM 0000000001FB0584 mod_ncparam_mp_in 1656 mod_ncparam.f90
oceanM 0000000000526FF5 read_phypar_ 179 read_phypar.f90
oceanM 0000000000428F6D inp_par_ 90 inp_par.f90
oceanM 000000000040BB0B ocean_control_mod 90 ocean_control.f90
oceanM 000000000040B202 MAIN__ 97 master.f90
oceanM 000000000040AF36 Unknown Unknown Unknown
libc.so.6 00002B8D27DF4CDD Unknown Unknown Unknown
oceanM 000000000040AE29 Unknown Unknown Unknown
iitmlogin3:/gpfs3/home/ibm/GOYAL/shivali/RKU/ROMS_kate_umaine $

Regards,
Shivali

[kate]

Something to check is your varinfo.dat file. Do you have all the cosine variables in there? I have a chunk starting with:

Code: Select all

!--------------UMAINE variables

[shivaligangwar]

Sir,

Please check my varinfo.dat.fr .. i have attached the same

[kate]

Try the file Apps/NEP/varinfo.dat or simply Apps/varinfo.dat, depending on which is in your version. It should have that UMAINE section which you don't have in yours.

You can look at the file umaine_var.h to see the list of variables, starting with NO3. Your file has NO3, plus the next one, NH4, but not some of the rest. All of them have to be in your varinfo.dat.

[shivaligangwar]

Thanks a lot Sir,
That helped me to resolve the issue ...

But now i am stucked in next step

forrtl: severe (408): fort: (8): Attempt to fetch from allocatable variable NBCCOUNT when it is not allocated

Image PC Routine Line Source
oceanG 0000000001D437BC read_phypar_ 1383 read_phypar.f90
oceanG 0000000001A7558D inp_par_ 90 inp_par.f90
oceanG 000000000040B8E5 ocean_control_mod 84 ocean_control.f90
oceanG 000000000040B18E MAIN__ 95 master.f90
oceanG 000000000040AF26 Unknown Unknown Unknown
libc.so.6 00002B7149C7FCDD Unknown Unknown Unknown
oceanG 000000000040AE19 Unknown Unknown Unknown
forrtl: warning (402): fort: (1): In call to LOAD_L, an array temporary was created for argument #4

[shivaligangwar]

Sir,

error is coming at location - in readphypar.f90
forrtl: severe (408): fort: (8): Attempt to fetch from allocatable variable NBCCOUNT when it is not allocated

Image PC Routine Line Source
oceanG 0000000001D438EC read_phypar_ 1384 read_phypar.f90
oceanG 0000000001A7558D inp_par_ 90 inp_par.f90
oceanG 000000000040B8E5 ocean_control_mod 84 ocean_control.f90


CASE ('BRYNAME')
label='BRY - lateral open boundary conditions'
Npts=load_s2d(Nval, Cval, line, label, ibcfile, igrid, &
& NBCcount, mFfiles, nBCfiles, BRY)

[kate]

Ah, now you need to check your ocean.in to see if it matches one of my ocean.in files. I hacked things to allow more than one boundary file and now you need:

Code: Select all

     NBCFILES == 1                          ! number of boundary files
     BRYNAME == ocean_bry.nc

The allocate statement for that thing happens once it knows how many boundary files you have:

Code: Select all

            CASE ('NBCFILES')
              Npts=load_i(Nval, Rval, Ngrids, nBCfiles)
              DO ng=1,Ngrids
                IF (nBCfiles(ng).le.0) THEN 
                  IF (Master) WRITE (out,260) 'NBCFILES', nBCfiles(ng), &
     &                            'Must be equal or greater than one.'
                  exit_flag=4
                  RETURN
                END IF
              END DO
              max_Ffiles=MAXVAL(nBCfiles)
              allocate ( BRY(max_Ffiles,Ngrids) )
              allocate ( BRYids(max_Ffiles,Ngrids) )
              allocate ( NBCcount(max_Ffiles,Ngrids) )
              BRYids(1:max_Ffiles,1:Ngrids)=-1
              NBCcount(1:max_Ffiles,1:Ngrids)=0

P.S. Please don't call me "sir". Most women don't like it.

[shivaligangwar]

Thanks a lot again Ma'm ...
Putting NBCFILES in ocean.in helped me to remove error i was getting.

[shivaligangwar]

Dear Ma'm,

I am getting following error -

forrtl: severe (193): Run-Time Check Failure. The variable 'biology_mod_mp_biology_tile_$MIDDCA' is being used without being defined
Image PC Routine Line Source
oceanG 0000000001684531 biology_mod_mp_bi 122 biology.f90
oceanG 00000000006B806B main3d_ 286 main3d.f90
oceanG 000000000040CFA4 ocean_control_mod 164 ocean_control.f90
oceanG 000000000040B70D MAIN__ 108 master.f90
oceanG 000000000040AF26 Unknown Unknown Unknown
libc.so.6 00002AC697508CDD Unknown Unknown Unknown
oceanG 000000000040AE19 Unknown Unknown Unknown

I checked biology.f90, i can see MIDDCA getting declared as real but couldnt see it getting defined anywhere

its getting used in line 867 -
==> Qsms32=apsilon(ng)*(gc3zz2+MORTc3)- MIDDCA

removing MIDDCA helped me to resolve the issue
==> Qsms32=apsilon(ng)*(gc3zz2+MORTc3)

Please guide me, whats the significance of MIDDCA.


Regards,
Shivali

[kate]

I'm afraid I don't know. I have asked Peng in an email and will get back to you if I hear (or he can post here perhaps).

[kate]

Peng got back to me right away with an updated version of that file.

[shivaligangwar]

Thanks a lot Ma'm.. My ROMS model ran successfully with new updated header bioUMAINE31.h

Regards,
Shivali