Segmentation fault building CICE6.6.2 with intel compilers on linux workstation

[apauling]

What version of the code are you using?
CICE6.6.2

Have you made any changes to files in the source tree?
I have attempted to create a machine port for my linux workstation using existing examples in the configuration/scripts/machines directory.

Describe every step you took leading up to the problem:
- Ran the default CICE configuration using the conda machine spec. Built and ran with no issues
- Tried to create a machine port with the Intel OneAPI 2025.0 compilers. Compiled netcdf-c and netcdf-fortran myself using these same compilers which both passed all tests
- Model compiles successfully with the new port. However, when running it reaches a segmentation fault in ice_transport_remap.F90 in the "horizontal_remap" routine just after it tries to read the forcing file JRA55_gx3_03hr_forcing_2005.nc. I have tried recompiling and running in debug mode but this didn't provide any more information that I could see.

If this is a port to a new machine: Please attach any files you added or changed for the machine port (e.g., config_compilers.xml, config_machines.xml, and config_batch.xml) and tell us the compiler version you are using on this machine.
Please attach any log files showing error messages or other useful information.

I have attached the machine port files and the run log.

Describe your problem or question:

Issue described above, any help would be much appreciated.

 

[apauling]

Sorry, files attached here

 

[dbailey]

We have had some issues with the intel compilers on some systems. Can you try the gnu compiler? Do you have threading on?

 

[apauling]

Hi Dave, thanks for the quick reply. I can compile and run successfully with the gnu compiler as part of the conda config. I don't have threading on. I am hoping to be able to use the intel compiler as we are hoping to eventually couple CICE to an existing regional model, and this currently uses intel. I did try compiling with an older intel version (2023.2.4, which still uses ifort/icc instead of ifx/icx), but ran into the exact same issue.

 

[apauling]

Ugh, sorry. I figured this out. I had "ulimit -s unlimited" in my cice.run script, but since this is a csh script it needed to be "limit stacksize unlimited". Model is running fine now. Thanks again.