Pythia8 no output when using condor cluster

Asked by Diana Krupova on 2019-03-25

Hello,
I'm trying to generate events with Pythia8. When I'm doing it locally, it works alright, I just needed to add path to MG5aMC_PY8_interface as PYTHONPATH environment variable in ~/.bashrc. When I then try to submit the jobs to cluster using "launch --cluster", it splits the event file, submits jobs, says all jobs are finished, but then: "Merging results from the split PY8 runs...
Fail to produce a pythia8 output. More info in .../Events/run_01/tag_1_pythia8.log".

In the .log file, I always get this error message:
===================================
 -> Pythia8 log file for run 9 <-
===================================
condor_exec.exe: line 2: ./MG5aMC_PY8_interface: No such file or directory

So it seems to me like condor cannot find the MG5aMC_PY8_interface even though it is installed in HEPTools and there is a path to it in my mg5_configuration.txt file. C

Thanks
Diana

Question information

Language:
English Edit question
Status:
Open
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Valentin Hirschi Edit question
Last query:
2019-03-28
Last reply:
2019-03-27
Diana Krupova (krupodia) said : #1

I forgot to mention that I'm using MG5_aMC v2.6.4.

Hi,

I do not reproduce this with my cluster.
But many parameters are different (this is a slurm cluster) In principle I should have access to a condor cluster soon but I do not have working access yet.

Now the code is not designed to work with any alteration to the PYTHONPATH.
Why do you need to do that? (what is the error if you do not do that)

It is expected that some feature are disabled if you do not specify the path to a MG5aMC code in the file
Cards/me5_configuration.txt, but Pythia8 should work independently of that (at least if you have specify the pythia8 path
inside that file)

Cheers,

Olivier

> On 26 Mar 2019, at 21:52, Diana Krupova <email address hidden> wrote:
>
> Question #679538 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/679538
>
> Diana Krupova gave more information on the question:
> I forgot to mention that I'm using MG5_aMC v2.6.4.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Diana Krupova (krupodia) said : #3

Hi Olivier,
I already had specified both pythia8 and MG5aMC_py8_interface paths in the configuration file.
I deleted the PYTHONPATH and tried to generate events with cluster again but I get exactly the same error.

Let me wait for my access to a condor cluster since this might be condor specific.

Cheers,

Olivier

> On 26 Mar 2019, at 23:57, Diana Krupova <email address hidden> wrote:
>
> Question #679538 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/679538
>
> Status: Answered => Open
>
> Diana Krupova is still having a problem:
> Hi Olivier,
> I already had specified both pythia8 and MG5aMC_py8_interface paths in the configuration file.
> I deleted the PYTHONPATH and tried to generate events with cluster again but I get exactly the same error.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Diana Krupova (krupodia) said : #5

Hi Olivier,
it works now, I just had to change the cluster temp path in me5_configuration file to the directory where I got the MG5 installed (even though it says this should affect condor).

So I'm generating the events now but I've got one more question, is there a way to change the number of jobs it splits the .lhe file for PY8 parallelization into? I've tried changing the cluster size to 500 in me5_configuration but it had no effect.
It's too slow now, as I'm generating 10M events with multi_run split into 1000 runs, no more than 5% of them was finished in the past 20 hours...

Can you help with this problem?

Provide an answer of your own, or ask Diana Krupova for more information if necessary.

To post a message you must log in.