MadGraph5_aMC@NLO

Error with pythia8 parallelization when matching

Asked by Jack Collins

When I run any set of processes requiring jet matching in the MG 2.5.1, I get the following error:

InvalidCmd : Pythia8 shower failed since it did not accept any event from the MG5aMC event file.

When I go the the folder Events/my_events/PY8_parallelization, there are many folders called split_n. For n > 0, the files split_n/PY8_log.txt look fine, and there is a hepmc file with the size I expect, so it looks like the showering worked fine for these. For n = 0, the hepmc file is empty and the split_0/PY8_log.txt file has:

 -------- LHA initialization information ------------

  beam kind energy pdfgrp pdfset
     A 2212 4000.000 0 247000
     B 2212 4000.000 0 247000

  Event weighting strategy = -4

  Processes, with strategy-dependent cross section info
  number xsec (pb) xerr (pb) xmax (pb)
       1 1.9336e-04 1.8440e-06 1.9898e-03
       2 2.7283e-05 3.5104e-07 1.9898e-03
       5 3.4526e-04 2.5803e-06 1.9898e-03
       6 1.9599e-04 2.2442e-06 1.9898e-03

 -------- End LHA initialization information --------
 PYTHIA Abort from Pythia::next: reached end of Les Houches Events File
WARNING in MG5aMC_PY8_interface.cc: Reached end of LHEF after 0 Normalisation will be decreased by-100%.

 *------- PYTHIA Event and Cross Section Statistics -------------------------------------------------------------*
 | |
 | Subprocess Code | Number of events | sigma +- delta |
 | | Tried Selected Accepted | (estimated) (mb) |
 | | | |
 |-----------------------------------------------------------------------------------------------------------------|
 | | | |
 | Les Houches User Process(es) 9999 | 0 0 0 | 0.000e+00 0.000e+00 |
 | | | |
 | sum | 0 0 0 | 0.000e+00 0.000e+00 |
 *------- End PYTHIA Event and Cross Section 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

This happens for any process I have tried to run where I am using jet matching, and does not happen if I do not attempt jet matching. I can't think how I can further diagnose the problem, it only seems to be a problem for split_0.

Question information

Language:
English Edit question
Status:
Answered
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Valentin Hirschi Edit question
Last query:
Last reply:
Revision history for this message
Valentin Hirschi (valentin-hirschi) said : 		#1

Can you confirm that you observe the same behavior in the new release v2.5.2?
If so, please send me a run_card (or list of commands for an example simple matched process) for which you observe the problem, as well as the number of events asked (if not the default 10k) so that i can try to reproduce it and investigate the problem.

Revision history for this message
Jack Collins (jhc296) said : 		#2

I run:

generate p p > z j, z > e+ e-
add process p p > z j j, z > e+ e-

In the run_card.dat I set xqcut = 20 and have all jet cuts (besides eta) set to 0. ickkw = 1.

I use the default pythia8_card, except I change QCUT from -1 to 30 (a separate problem I am having is that if I leave it as -1, which the file says means it should choose it automatically, then I get error messages saying -1.0 is too small a value for this parameter and the run fails. It seems like the automatic choosing is not working like it's supposed to).

This is sufficient for me to get the error I described.

Revision history for this message
Jack Collins (jhc296) said : 		#3

I run:

generate p p > z j, z > e+ e-
add process p p > z j j, z > e+ e-

In the run_card.dat I set xqcut = 20 and have all jet cuts (besides eta) set to 0. ickkw = 1.

I use the default pythia8_card, except I change QCUT from -1 to 30 (a separate problem I am having is that if I leave it as -1, which the file says means it should choose it automatically, then I get error messages saying -1.0 is too small a value for this parameter and the run fails. It seems like the automatic choosing is not working like it's supposed to).

This is sufficient for me to get the error I described.

Revision history for this message
Valentin Hirschi (valentin-hirschi) said : 		#4

Thank you for your feedback.

 I am aware of the QCUT = -1 issue and it will be fixed for the next
release.

As for the issue in the parallelization, could you tar-zip the directory
Events/my_events/PY8_parallelization (removing all hepmc file in it so as
it's reasonably light) and attach it to this launchpad thread?

I will look into it.

On Thu, Dec 15, 2016 at 3:12 PM, Jack Collins <
<email address hidden>> wrote:

> Question #404633 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/404633
>
> Jack Collins gave more information on the question:
> I run:
>
> generate p p > z j, z > e+ e-
> add process p p > z j j, z > e+ e-
>
> In the run_card.dat I set xqcut = 20 and have all jet cuts (besides
> eta) set to 0. ickkw = 1.
>
> I use the default pythia8_card, except I change QCUT from -1 to 30 (a
> separate problem I am having is that if I leave it as -1, which the file
> says means it should choose it automatically, then I get error messages
> saying -1.0 is too small a value for this parameter and the run fails.
> It seems like the automatic choosing is not working like it's supposed
> to).
>
> This is sufficient for me to get the error I described.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.
>

--
Valentin

Revision history for this message
Valentin Hirschi (valentin-hirschi) said : 		#5

I just realized that your process started with a light jet in your lowest multiplicity sample, i.e.:

generate p p > z j, z > e+ e-
add process p p > z j j, z > e+ e-

I believe this is not allowed in MLM-type of merging. You should therefore either consider p p > z, z > e+ e- as well, or use the alternative merging scheme CKKW-L now also available in MG5_aMC (keep ickkw=0 and set either the ktdurham or ptlund cut to a non-zero value in the run_card. You will also need to define the lowest multiplicity process in the Pythia8 shower card in this case, and you can find more information about this in the PY8 manual.
Contact Pythia8 authors for further details on the CKKW-L merging scheme.

Can you help with this problem?

Provide an answer of your own, or ask Jack Collins for more information if necessary.

To post a message you must log in.