ICE_AGRIF failed in sette

in 1_ocean.output :

==> Child Bathymetry is NOT correct near boundaries.

changed the description

assigned to @jchanut

added 1 deleted label

should be corrected now

I still have an issue with ICE_AGRIF in debug mode at rev d47e9a4e:

forrtl: error (65): floating invalid
Image              PC                Routine            Line        Source
nemo               0000000004E548EE  Unknown               Unknown  Unknown
libpthread-2.17.s  00002ABEF86E7630  Unknown               Unknown  Unknown
nemo               0000000002FDDECA  icesbc_mp_sub_loo         650  icesbc.f90
nemo               0000000002FCF7FC  icesbc_mp_ice_flx         505  icesbc.f90
nemo               0000000002FB644E  icesbc_mp_sub_loo         315  icesbc.f90
nemo               0000000002FB2136  icesbc_mp_ice_sbc         201  icesbc.f90
nemo               0000000000CD7691  icestp_mp_sub_loo         274  icestp.f90
nemo               0000000000CD2A64  icestp_mp_ice_stp         131  icestp.f90
nemo               00000000004EE105  sbcmod_mp_sub_loo         767  sbcmod.f90
nemo               00000000004D986B  sbcmod_mp_sbc_            510  sbcmod.f90
nemo               00000000006004DE  step_mp_sub_loop_         176  step.f90
nemo               00000000005FD73A  step_mp_stp_               88  step.f90
nemo               0000000000450822  nemogcm_mp_sub_lo         197  nemogcm.f90
nemo               000000000044E877  nemogcm_mp_nemo_g         110  nemogcm.f90
nemo               000000000044DC01  MAIN__                     45  nemo.f90
nemo               000000000044DB8E  Unknown               Unknown  Unknown
libc-2.17.so       00002ABEF8C18555  __libc_start_main     Unknown  Unknown
nemo               000000000044DAA9  Unknown               Unknown  Unknown

When and where was this supposedly fixed? I can't see a relevant commit?

Commit 7605d1a4 (a commit that is embedded in a sea-ice commit). I just changed the namelist, so that bathy is not checked in Agrif. We do not care for this test case since it is a standalone case

The problem comes from the use of external mbku/mbkv/mbkf arrays I introduced on svn before turning to Git. More precisely, it comes from dommsk.F90, lines 163-165. This does not work for the (extreme) case of jpk=2. This explains also the problems encountered for SWG test case.

mentioned in commit eece361f

mentioned in merge request !11 (merged)

mentioned in commit 30abfe8a

closed

reopened

mentioned in commit fff940c9

closed with commit eece361f

closed with merge request !11 (merged)

changed milestone to %Release 4.2