MadGraph5_aMC@NLO

Too few events after reweighting

Asked by Chris Brust

Hi,

   I'm attempting to generate and decay 50k events in a FR-created model. Things seem to go fine during event generation; however, after the reweighting step, I only have between 5k and 20k events, depending on the parameters in the model. As far as I can tell, no particle that I'm attempting to decay has 0 width, and the problem goes away if I only decay the Ws leptonically, or if I don't decay the stops (which seems quite odd!). I'm using the run_card provided in the Template directory with a few minor changes (root(s) = 7 TeV, number of events), which I can post if desired. The proc_card for the model in question is the following:

import model uddrpv
# Define multiparticle labels
define p = g u c d s u~ c~ d~ s~
define j = g u c d s u~ c~ d~ s~ b b~
define l+ = e+ mu+ ta+
define l- = e- mu- ta-
define vl = ve vm vt
define vl~ = ve~ vm~ vt~
# Specify process(es) to run
generate p p > b1 b1~, (b1 > w- t1, w- > j j, t1 > j j), (b1~ > w+ t1~, w+ > j j, t1~ > j j) QCD=4 QED=6
add process p p > b1 b1~, (b1 > w- t1, w- > l- vl~, t1 > j j), (b1~ > w+ t1~, w+ > j j, t1~ > j j) QCD=4 QED=6
add process p p > b1 b1~, (b1 > w- t1, w- > j j, t1 > j j), (b1~ > w+ t1~, w+ > l+ vl, t1~ > j j) QCD=4 QED=6
add process p p > b1 b1~, (b1 > w- t1, w- > l- vl~, t1 > j j), (b1~ > w+ t1~, w+ > l+ vl, t1~ > j j) QCD=4 QED=6
# Output processes to MadEvent directory
output -f

The diagrams appear to be what I expect. Any help would be greatly appreciated. Thank you!

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Johan Alwall
Solved:
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

The problem might be related to the mass spectrum. By using this syntax, you force the particle to be onshell.
(i.e that
| m^* - m | < bw_cutoff * width
 (where bw_cutoff is defined in the run_card.dat)

Could you send to me (olivier.mattelaer AT uclouvain.be) the model and the cards that you use, such that I can take a look.

Cheers,

Olivier

Revision history for this message
Chris Brust (cpbrust) said : 		#2

Olivier,

   The widths in my model are all relatively small, and I haven't changed bwcutoff from its default 15. Regardless, I've emailed you the information requested. Thank you! Chris

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#3

Hi Chris,

Just to keep you up-to-date with our progress.
I have found that a segmentation fault was responsible to a silent
crash forbidding the code to pass
by the refine steps (which allows to reach the requirements of events).

I'm going to discuss more into details with Johan, and this should be
solve in the next version.

Thanks a lot,

olivier

On Apr 10, 2012, at 3:35 PM, Chris Brust wrote:

> Question #193105 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/193105
>
> Status: Needs information => Open
>
> Chris Brust gave more information on the question:
> Olivier,
>
> The widths in my model are all relatively small, and I haven't
> changed bwcutoff from its default 15. Regardless, I've emailed you the
> information requested. Thank you! Chris
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Revision history for this message
Chris Brust (cpbrust) said : 		#4

Thank you very much! In the meantime, do you suspect that the ~10k events that are generated are weighted correctly and otherwise behave properly, or should I wait for an update before attempting to generate these events? Thank you, Chris

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#5

Hi Chris,

They should be fine. No problem with those.

Cheers,

Olivier

On 12-avr.-12, at 17:00, Chris Brust wrote:

> Question #193105 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/193105
>
> Chris Brust posted a new comment:
> Thank you very much! In the meantime, do you suspect that the ~10k
> events that are generated are weighted correctly and otherwise behave
> properly, or should I wait for an update before attempting to generate
> these events? Thank you, Chris
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Revision history for this message
Best Johan Alwall (johan-alwall) said : 		#6

Hello Chris,

It seems that the problem is due to a bug in v. 1.4.3 that was fixed in v. 1.4.4 (see the Change log/UpdateNotes:

JA: Fixed buffer overflow in gen_ximprove when number of
                      configs > number of diagrams due to competing resonances
                      (introduced in v. 1.4.3).
)

Can you please try running with v. 1.4.4 to make sure that the problem is fixed there?

Many thanks,
Johan

Revision history for this message
Chris Brust (cpbrust) said : 		#7

Will do; I'll get back to you once the run has completed. Chris

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#8

Just to confirm that my version was also not up-to-date and that I was
running under 1.4.3

Cheers,

Olivier

On Apr 13, 2012, at 8:01 AM, Chris Brust wrote:

> Question #193105 on MadGraph5 changed:
> https://answers.launchpad.net/madgraph5/+question/193105
>
> Chris Brust posted a new comment:
> Will do; I'll get back to you once the run has completed. Chris
>
> --
> You received this question notification because you are a member of
> MadTeam, which is an answer contact for MadGraph5.

Revision history for this message
Chris Brust (cpbrust) said : 		#9

Using 1.4.4 did indeed solve my problem. Thank you! Chris

Revision history for this message
Chris Brust (cpbrust) said : 		#10

Thanks Johan Alwall, that solved my question.