search

FESOM / fesom2

Multi-resolution ocean general circulation model.
http://fesom.de/
GNU General Public License v3.0
45 stars 46 forks source link

Model blowup on ecmwf-atos for TCO319-DART configuration, mzapponi_refactoring4awicm3.1 branch #611

Open tsemmler05 opened 3 days ago

tsemmler05 commented 3 days ago

When running above model configuration with git commit bcbb1172, the coupled model blows up after a number of time steps. The number of time steps depends on if it is a cold start or a restart run. If it is a restart run, it also depends on which restart files are used. eta_z is from one time step to the next out of the range -10 / +10; from one to the next time step the temperature drops from a reasonable value (for example 8 or 15 C) to a large negative value (around -25 C).

When I change to the git commit e2f45f07, the one that is running without any problems on DKRZ levante, I get a seg fault straight away:

FESOM start iteration before the barrier... FESOM start iteration after the barrier...

^[[32m____^[[0m ^[[7;32m --> FESOM STARTS TIME LOOP ^[[0m file : namelist.io open ok io_listsize will be changed from 100 to 39 ! adding I/O stream 2D for sst [ac1-3055:2759318:0:2759318] Caught signal 11 (Segmentation fault: address not mapped to object at address 0x10)

JanStreffing commented 3 days ago

Hey Tido, could you compile with debug flags to change the error into something more interpretable? To do that you need to change a bit the src/CMakeLists.txt . Which compile are you using?