Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

execute cesm.exe

Vru

Vru
New Member
Hello

I am trying to run cesm without using case.submit (or .case.run), I mean directly executing cesm.exe, but that does not work at all.

My porting is otherwise correct since a "normal" ./case.submit is successful, and I do not understand what is missing.

All the needed files seem to be present in the rundir, and still executing cesm.exe fails !

I guess that there must be environment variables missing, but what is it that the scripts do that I could not reproduce step by step?

Thank you in advance for you help
 

jedwards

CSEG and Liaisons
Staff member
Why? You need to source the file in your case directory .env_mach_specific.sh (or csh)
to set up the modules and environment variables.

You can view what the scripts do with the preview_run script.
 

Vru

Vru
New Member
There is in the case directory an env_mach_specific.xml, how do I generate this env_mach_specific.sh?
 

Vru

Vru
New Member
There must be something else missing because sourcing .env_mach_specific.sh only added export KMP_STACKSIZE=64M, and the execution still fails the same when directly calling cesm.exe (whereas it is successful when using case.submit or python .case.run)

For some reason in debug mode the crash happens a bit farther:

Program received signal SIGFPE: Floating-point exception - erroneous arithmetic operation.

Backtrace for this error:
#0 0x7fd30aa293ff in ???
#1 0x1071527 in __micro_mg2_0_MOD_micro_mg_tend
 

jedwards

CSEG and Liaisons
Staff member
You never answered the question - why? Why are you attempting to bypass the cesm tools and methods? Before you post any further errors or backtraces please make sure that they also occur using the approved build and submit procedures.
 

Vru

Vru
New Member
I am trying to run cesm inside a singularity container, which requires starting the application by calling the mpi launcher from the host (i.e., mpirun -n 16 singularity ... cesm.exe, since obviously that does not work with ./case.submit)

The execution is successful with the approved build and submit procedures when not using containers

I do not understand how to use the standard scripts with singularity
 

jedwards

CSEG and Liaisons
Staff member
Why does this obviously not work from case.submit? You can edit the contents of the mpirun line in config_machines.xml and view it with preview_run, I see no reason case.submit would not work in this situation.
 

Vru

Vru
New Member
My problem is that I did not manage to configure the appropriate config_batch.xml and config_machines.xml for use with slurm + singularity (mpirun -np $TOTALPES singularity ...)

It may be possible but I could not figure out how to do that, all examples I found are for use without container or with container on a single node
 
Top