New URL for NEMO forge!   http://forge.nemo-ocean.eu

Since March 2022 along with NEMO 4.2 release, the code development moved to a self-hosted GitLab.
This present forge is now archived and remained online for history.
#2639 (Compiling DOME on ARCHER2 fails) – NEMO

Opened 3 years ago

Closed 3 years ago

#2639 closed Bug (fixed)

Compiling DOME on ARCHER2 fails

Reported by: jamesharle Owned by: systeam
Priority: low Milestone:
Component: USR Version: trunk
Severity: minor Keywords: usr_def,
Cc:

Description

Context

When compiling the DOME test case on UK ARCHER2 HPC it fails with the following message:

IMPLICIT NONE is specified in the host scope, therefore an explicit type must be specified for function "MIG0_OLDCMP".

Analysis

It doesn't appear mig0_oldcmp is defined in the code base, but I'm not 100% familiar with this test case. Looking at BENCH for a comparison I presume this variable should be mig0?

Fix

  • usrdef_hgr.F90

     
    8888#endif 
    8989      DO_2D( 1, 1, 1, 1 ) 
    90          zti = REAL( mig0_oldcmp(ji) - 1, wp )   ! start at i=0 in the global grid without halos 
    91          ztj = REAL( mjg0_oldcmp(jj) - 1, wp )   ! start at j=0 in the global grid without halos 
     90         zti = REAL( mig0(ji) - 1, wp )   ! start at i=0 in the global grid without halos 
     91         ztj = REAL( mjg0(jj) - 1, wp )   ! start at j=0 in the global grid without halos 
    9292         plamt(ji,jj) = rn_dx * 1.e-3 * ( zlam0 + zti ) 
    9393         plamu(ji,jj) = rn_dx * 1.e-3 * ( zlam0 + zti + 0.5_wp ) 

Commit History (0)

(No commits)

Change History (3)

comment:1 Changed 3 years ago by smasson

Yes, I introduced mig0_oldcmp when we were testing the suppression of the MPI halos from the input/output files. This function was allowing me to get exactly the same results and validate the modifications.
It was a temporary feature and I removed it in [14223].
tests/DOME/MY_SRC/usrdef_hgr.F90 was introduced later at [14254] with this "old" feature.

Note that replacing mig0_oldcmp by mig0 generates a shift of 1. I think that the modifications of DOME should follow what was done in BENCH for example to make sure the generated grid is still centered around the exact same point. see
https://forge.ipsl.jussieu.fr/nemo/changeset/14223/NEMO/trunk/tests/BENCH/MY_SRC/usrdef_hgr.F90

comment:2 Changed 3 years ago by jchanut

My mistake. DOME was tested with a trunk version right before Seb changes @1423. I certainly forgot to checkout the code before updating...
An other replacement of mig0_oldcmp can be found here:
https://forge.ipsl.jussieu.fr/nemo/changeset?sfp_email=&sfph_mail=&reponame=&new=14223%40NEMO%2Ftrunk%2Ftests%2FVORTEX%2FMY_SRC%2Fusrdef_hgr.F90&old=13295%40NEMO%2Ftrunk%2Ftests%2FVORTEX%2FMY_SRC%2Fusrdef_hgr.F90

comment:3 Changed 3 years ago by jchanut

  • Resolution set to fixed
  • Status changed from new to closed

Repaired @ r14976

Note: See TracTickets for help on using tickets.