Core dumping issues

[scheah86]

Issue summary

Hi there. I've experienced core dump issues while processing 3.7 million compounds on the supercomputer. I submitted a total of 1,488 jobs and received the same number in return. However, I've noticed that several of my text files indicate a "slurmstep" error. Could you provide guidance on resolving this issue? Thank you.

Steps to reproduce

1. Log into a supercomputer account
2. Run GNINA script

Your system configuration

Operating system:
Compiler:
CUDA version (if applicable):
CUDNN version (if applicable):
Python version:

[AquifersBSIM]

I am experiencing similar issues as scheah86 but with slightly different outcome. What happened in my case was that a job was submitted for molecular docking, and the expected number of molecules was supposed to be 1200, instead a lower number was produced, let's say 600. This is definitely puzzling because it was presumed that when core dumping happened, the immediate action was to restart the molecular docking. With the immediate restart, the molecular docking proceeds without core dumping.

Assuming that it is a potential bug in the code, and I am keen to understand why does it occur, let's say, 10 in 80 jobs. Upon attempting to resolve the 10 jobs, some were executed successfully, others still generates core dumps.

[scheah86]

Hi there. I would like to update about this issue. I tried rerunning the screening process as suggested by the commenter above. I realised that the core dumping issue still occurs but happens at random, like the example below.

May I know if it is possible to resolve this issue? Thank you.