Problem with madevent and scan files

Asked by Luca Marzola

Hello,

I am trying to feed a scan file to madevent after having create the folder for the process via mg5.

The scan file looks like this:

launch

# Analysis setup

# Collider setup
set ebeam1 7000
set ebeam2 7000

set run_tag = 'point_1'
set MSCp 1000.0
set MH2 25

launch
set run_tag = 'point_2'
set MSCp 1000.0
set MH2 250

launch
set run_tag = 'point_3'
set MSCp 2000.0
set MH2 25

launch
set run_tag = 'point_4'
set MSCp 2000.0
set MH2 250

launch
set run_tag = 'point_5'
set MSCp 3000.0
set MH2 25
....

With this, madevent computes the cross section for the first point only and then exits (no errors reported). If I swap "launch" - > "generate_events", instead, it creates a reweight card and computes the last point only. I can see however that it has set the parameters for the previous runs too, but the cross sections are not reported.

I intended to have it compute the cross sections a bunch of times and then read them out with the print_results command.

How should I change my file so that madevent doesn't quit after the first run?

Thank you for your help!

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Luca Marzola
Solved:
Last query:
Last reply:
Revision history for this message
Luca Marzola (lucamarzola) said (last edit ):
#1

So, I investigated some more.

I believe the culprit was lhapdf: it was silently causing python to crash. I uninstalled it and everything works as expected.

There's indeed something funny going on with lhapdf6 and gcc 11/ m1 processors. I originally installed it via madgraph (after installing wget via homebrew, as it is required to download pdfs) but still there must be something not working.

If I download lhapdf from the official website and try to compile it, I get some errors due to gcc 11 not having some locale headers.
Anyway, looks like madgraph itself is fine.