MadGraph5_aMC@NLO

event generation fails with MLM matching

Asked by Samuel Homiller

Hi Everyone,

I'm attempting to simulate the process p p > b b~ a a as well as the process with an extra jet using MLM matching along with Pythia6. The simulation seems to work fine with both processes if I turn off matching (set ickkw 0), but when I turn on matching (ickkw 1, xqcut = 25.0), the process fails near the end of event generation. I had this simulation working some time ago, but I'm afraid that I've since updated madgraph and that something has changed since then, as re-running my old setup failed in the same way. I've attached the output log for the failed run below. Thanks very much for your help.

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
Last query:
Last reply:

This question was originally filed as bug #1682824.

Revision history for this message
Samuel Homiller (shomiller) said : 		#1
Revision history for this message
Best Olivier Mattelaer (olivier-mattelaer) said : 		#2

Hi,

Here is a patch that fix such problem:

=== modified file 'madgraph/various/lhe_parser.py'
--- madgraph/various/lhe_parser.py 2017-03-30 15:02:16 +0000
+++ madgraph/various/lhe_parser.py 2017-04-24 13:52:52 +0000
@@ -1263,7 +1263,7 @@

         if not hasattr(Event, 'loweight_pattern'):
             Event.loweight_pattern = re.compile('''<rscale>\s*(?P<nqcd>\d+)\s+(?P<ren_scale>[\d.e+-]+)\s*</rscale>\s*\n\s*
- <asrwt>\s*(?P<asrwt>[\s\d.+-]+)\s*</asrwt>\s*\n\s*
+ <asrwt>\s*(?P<asrwt>[\s\d.+-e]+)\s*</asrwt>\s*\n\s*
                                     <pdfrwt\s+beam=["']?1["']?\>\s*(?P<beam1>[\s\d.e+-]*)\s*</pdfrwt>\s*\n\s*
                                     <pdfrwt\s+beam=["']?2["']?\>\s*(?P<beam2>[\s\d.e+-]*)\s*</pdfrwt>\s*\n\s*
                                     <totfact>\s*(?P<totfact>[\d.e+-]*)\s*</totfact>

An other method is to switch off the systematics computation (and/or using SysCalc for that).
One way is to set use_syst on False in your run_card.

Cheers,

Olivier

Revision history for this message
Samuel Homiller (shomiller) said : 		#3

Thanks Olivier Mattelaer, that solved my question.