Comment 12 for bug 1674341

So there seems to be two different issues in this thread:

----------
1) Sometimes, the split event files show "empty events" which crash Pythia8.
I cannot reproduce this issue, even when starting from the 'unweighted_events.lhe' file provided by nicolas.neill.
So, at this stage, we cannot bring any further help. If you encounter similar issue (i.e. not just Pythia8 crashing, but crashing because it didn't accept any event, and then looking into the source .lhe file fed to PY8, you see empty events there), then please go to the file

<MG5_aMC_root_dir>/madgraph/interface/madevent_interface.py, around line 4106,

change:

n_splits = lhe_file.split(partition=partition, cwd=parallelization_dir, zip=True)

into:

misc.sprint("Original evt file:", pjoin(self.me_dir,'Events',self.run_name,PY8_Card.subruns[0]['Beams:LHEF']))
misc.sprint("Partitions asked for":, partition)
n_splits = lhe_file.split(partition=partition, cwd=parallelization_dir, zip=True)

And then send us the event file pointed at by the first printout line and what the 'partition' variable is. This should allow us to straight-forwardly reproduce the issue.

---------
2) The second issue 'Nelson Vanegas-Arbelaez' is having is probably completely unrelated and happens both with and without parallelization.
In this case it is probably an issue during Pythia8 run itself.
I suggest you look inside the Pythia8 log file in:

NewBckgndWZ/Events/run_02

And going to the end of it, it is likely that you will find an error report from PY8. If not, then forward this log by attaching it here to this thread.
If you find an error however, and you believe it is not directly related to the way MG5_aMC produced the event file, then directly address this error to Pythia8 authors.