PY8 shower parallelization potential bug

The parallelization is unfortunately not handled by Pythia8 but MG5_aMC (it
should be the other way around but ok).
So it is MG5_aMC that split up the event files and has to sew them back
together, not Pythia8.

So you are saying that in your .../Events/run_xxx/tag_...._.hepmc.gz file
you only have 500 events and not 10K?
Please give me the exact path of the HEPMC event file that has only 500
events when you were expecting 10K.

On Wed, Jul 12, 2017 at 11:08 PM, Olivier Mattelaer <
<email address hidden>> wrote:

> Question #647792 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/647792
>
> Assignee: None => Valentin Hirschi
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.
>

--
Valentin

Thanks for the reply

/scratch/emc5ud/tt2j/Events/run_01/tag_1_pythia8_events.hepmc.gz (503 events)

/scratch/emc5ud/tt2j/Events/run_01/unweighted_events.lhe.gz (10K events)

Could you try the following two things:

a)
Turn off parallelization and check if you still get only 503 events. You
can do so like this:

cd /scratch/emc5ud
./bin/mg5
MG5_aMC> launch tt2j -i
tt2j > set nb_core 1
tt2j > shower pythia8 run_01
[... shower should start with parallelization off (on job, no event file
splitting)....]

b)
*After step a*, If you still only have ~500 events, something fishy
happened directly at the level of Pythia8. To better understand this, run
the shower by hand so as to directly see its progress:

cd /scratch/emc5ud/tt2j/Events/run_01
./run_shower.sh

Pay attention to Pythia8 warnings, probably indicating the reason why so
many events got rejected.

Cheers,

On Wed, Jul 19, 2017 at 11:23 PM, Eric Culbertson <
<email address hidden>> wrote:

> Question #647792 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/647792
>
> Status: Answered => Open
>
> Eric Culbertson is still having a problem:
> Thanks for the reply
>
> /scratch/emc5ud/tt2j/Events/run_01/tag_1_pythia8_events.hepmc.gz (503
> events)
>
> /scratch/emc5ud/tt2j/Events/run_01/unweighted_events.lhe.gz (10K events)
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.
>

--
Valentin

Dear Valentin

I have the same problem and I used step a) and b) as you told. It did not work finally.

---------------------------------
./run_shower.sh :

 PYTHIA Abort from Pythia::init: Les Houches Event File not found
Counting events in input LHEF, please wait.
Found 0 events in input LHEF.

Start generating events

 *------- PYTHIA Event and Cross Section Statistics -------------------------------------------------------------*
 | |
 | Subprocess Code | Number of events | sigma +- delta |
 | | Tried Selected Accepted | (estimated) (mb) |
 | | | |
 |-----------------------------------------------------------------------------------------------------------------|
 | | | |
 | | | |
 | sum | 0 0 0 | 0.000e+00 0.000e+00 |
 | |
 *------- End PYTHIA Event and Cross Section Statistics ----------------------------------------------------------*

 *------- PYTHIA Error and Warning Messages Statistics ----------------------------------------------------------*
 | |
 | times message |
 | |
 | 2 Abort from Pythia::init: Les Houches Event File not found |
 | |
 *------- End PYTHIA Error and Warning Messages Statistics ------------------------------------------------------*

 Contribution of sample 0 to the inclusive cross section : 0.00000000e+00 +- 0.00000000e+00

Inclusive cross section: 0.00000000e+00 +- 0.00000000e+00 mb

---------------------------------

In continue of previous message:

At first, I used pythia8 and delphes in the shell of Madgraph. For 100 events, there was 62 events in the final root file.
After that, I ran step a) and b) as explained in the previous message.

Please create your own bug report since this is something totally different.