With a python script found on GitHub (mep.py, revision 2019, which should build cube files for electrostatic potential from ORCA computations), code


at a mainframe raised error:

Sorry, while the code referes correctly to the mainframe, the error message refers from a different session on my laptop. The true error message from the mainframe (which tells the same issue) was
ORCA ELECTROSTATIC POTENTIAL GENERATION
------------------------------------------------------------------------------

GBW file ... qmmm_0.input.gbw
File with points to evaluate V(r) ... qmmm_0.input_mep.inp
Output file ... qmmm_0.input_mep.out

Reading the GBW file ... done
Reading the positions ... done (512000 positions)
Traceback (most recent call last):
File "mep.py", line 98, in <module>
basename + "_mep.inp", basename + "_mep.out"])
File "/cineca/prod/opt/compilers/python/3.6.4/none/lib/python3.6/subprocess.py", line 291, in check_call
raise CalledProcessError(retcode, cmd)
subprocess.CalledProcessError: Command '['/gpfs/work/IscrC_agelA/RNA-agelA_7MB640_sphere20/QM-MM_crystal_orca_paired/0/orca_vpot', 'qmmm_0.input.gbw', 'qmmm_0.input.scfp', 'qmmm_0.input_mep.inp', 'qmmm_0.input_mep.out']' died with <Signals.SIGKILL: 9>.
I also tried with python2.6 on my laptop getting the same error. The path is correct.

the error message is obscure to me, grateful for advice
chiendarret

The path is correct, of course, otherwise you would not get a segmentation violation, but an error to the effect that it can't be found.

Your problem is that /home/francesco/tmp2/orca_vpot crashes. Since I know exactly nothing about this program, this is all I can say.

You could try to launch the command from a shell prompt. Perhaps it has a verbose or debug flag that you can add. Perhaps it's a script, so that you can look into it. In any case, to understand why mep.py doesn't work, you need to find out why /home/francesco/tmp2/orca_vpot crashes on that computer.

EDIT: The Orca web site states
I am sure you are in better hands there.

orca_vpot was used as a standalone to generate, from ORCA QM computations, a list of partial atomic charges.

mep.py was asked now to use those charges to build the electrostatic maps.

Therefore, I suspect that the sigkill does not arise from problems related to orca_vpot, which was the reason why I posted here.

Anyway, I'll follow your advice to pose the question to ORCA.

By the way, I had to revise mep.py in order that it does not die immediately

thanks

yes, it is not related to python at all. I would say the environment is not prepared correctly. But it is a guess only.

I had to revise mep.py, and likely not exhaustively. For example, the original 2019 version of mep.py has code
whereby mep.py dies immediately with python3. mep.py did not work even with python2.6, so that I revised next there

and also elsewhere. That allowed mep.py to arrive at the crash that I posted previously.

Unfortunately, the ORCA forum is very little visited, and mep.py is not offered by ORCA officially.
Other (rare) free software to build electrostatic density maps also suffer from the problems of different python versions, and, unfortunately, as a biochemist, I lacked time to become a python expert. Such maps are of primary importance and commercial software does the job, but I have none.

All that merely to give a reason why I posted here, but I understand that one should exactly know what the code is asking to orca_vpot, which is not clear from the code.
thanks

yes, again that is not related to the segmentation fault at all. You can try to execute the command without python (and constructed by python):
And you will see if it dies too.
Either the environment is incorrect or the input files contain something "strange"

Behind a Segmentation Fault I am unable to see. I could suspect that 512000 positions are too much for the code, usually used for small molecules, but it is a mere suspicion. I tried to get a verbose output with "-v" before the out file but at no avail. I hope that ORCA will reply. It looks like a sleeping forum.

thanks