Problem when updated to latest Madgraph version

Asked by Animesh Chatterjee

I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.

Question information

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

Hi,

Thanks for the information,
Could you give us all the detailed needed to reproduce such problem? Without those it is impossible to comment.

Cheers,

Olivier

PS: i.e. I would need to have the full banner of the generation and/or the proc_card, the param_card, the run_card and potentially the model if this is not one of those that can be auto-downloaed in 2.6.1

> On 26 Jan 2018, at 23:52, Animesh Chatterjee <email address hidden> wrote:
>
> New question #663643 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/663643
>
> I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Animesh Chatterjee (animesh018) said :
#2

Hi,

I am generating fermionic dark matter events using p p > chi chi~ , where I consider fixed target experiment. There are two problems which I see now

  1. In the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one when I assign one of the beam energy as zero, I am getting error on phase space (earlier version was MG5_aMC_v2_2_3 and new version is MG5_aMC_v2_6_1), I am using the same param_card and run_card for both the version, same model as well.
  2. Theta distribution of the fermionic dark matter (for same parameters, same model and all other variables same ) for the earlier and latest version re not same.

I have attached the proc_card, the param_card, the run_card and the model herewith. Please let me know if you need more information.

Thank you for your help

Animesh

Regards
Animesh Chatterjee
Postdoctoral Research Fellow
Department of Physics
University of Texas at Arlington
________________________________
From: <email address hidden> <email address hidden> on behalf of Olivier Mattelaer <email address hidden>
Sent: Friday, January 26, 2018 4:57:33 PM
To: Chatterjee, Animesh
Subject: Re: [Question #663643]: Problem when updated to latest Madgraph version

Your question #663643 on MadGraph5_aMC@NLO changed:
https://answers.launchpad.net/mg5amcnlo/+question/663643
Question #663643 : Questions : MadGraph5_aMC@NLO<https://answers.launchpad.net/mg5amcnlo/+question/663643>
answers.launchpad.net
I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.

    Status: Open => Answered

Olivier Mattelaer proposed the following answer:
Hi,

Thanks for the information,
Could you give us all the detailed needed to reproduce such problem? Without those it is impossible to comment.

Cheers,

Olivier

PS: i.e. I would need to have the full banner of the generation and/or
the proc_card, the param_card, the run_card and potentially the model if
this is not one of those that can be auto-downloaed in 2.6.1

> On 26 Jan 2018, at 23:52, Animesh Chatterjee <email address hidden> wrote:
>
> New question #663643 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/663643
Question #663643 : Questions : MadGraph5_aMC@NLO<https://answers.launchpad.net/mg5amcnlo/+question/663643>
answers.launchpad.net
I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.

>
> I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

--
If this answers your question, please go to the following page to let us
know that it is solved:
https://answers.launchpad.net/mg5amcnlo/+question/663643/+confirm?answer_id=0

If you still need help, you can reply to this email or go to the
following page to enter your feedback:
https://answers.launchpad.net/mg5amcnlo/+question/663643
Question #663643 : Questions : MadGraph5_aMC@NLO<https://answers.launchpad.net/mg5amcnlo/+question/663643>
answers.launchpad.net
I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.

You received this question notification because you asked the question.

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

Attachment does not go trough the question on launchpad.
If you want to attach document, please create a bug report.

Cheers,

Olivier

> On 9 Feb 2018, at 18:43, Animesh Chatterjee <email address hidden> wrote:
>
> Question #663643 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/663643
>
> Status: Answered => Open
>
> Animesh Chatterjee is still having a problem:
> Hi,
>
>
> I am generating fermionic dark matter events using p p > chi chi~ , where I consider fixed target experiment. There are two problems which I see now
>
>
> 1. In the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one when I assign one of the beam energy as zero, I am getting error on phase space (earlier version was MG5_aMC_v2_2_3 and new version is MG5_aMC_v2_6_1), I am using the same param_card and run_card for both the version, same model as well.
> 2. Theta distribution of the fermionic dark matter (for same parameters, same model and all other variables same ) for the earlier and latest version re not same.
>
>
> I have attached the proc_card, the param_card, the run_card and the model herewith. Please let me know if you need more information.
>
>
> Thank you for your help
>
> Animesh
>
>
> Regards
> Animesh Chatterjee
> Postdoctoral Research Fellow
> Department of Physics
> University of Texas at Arlington
> ________________________________
> From: <email address hidden> <email address hidden> on behalf of Olivier Mattelaer <email address hidden>
> Sent: Friday, January 26, 2018 4:57:33 PM
> To: Chatterjee, Animesh
> Subject: Re: [Question #663643]: Problem when updated to latest Madgraph version
>
> Your question #663643 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/663643
> Question #663643 : Questions : MadGraph5_aMC@NLO<https://answers.launchpad.net/mg5amcnlo/+question/663643>
> answers.launchpad.net
> I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.
>
>
>
> Status: Open => Answered
>
> Olivier Mattelaer proposed the following answer:
> Hi,
>
> Thanks for the information,
> Could you give us all the detailed needed to reproduce such problem? Without those it is impossible to comment.
>
> Cheers,
>
> Olivier
>
> PS: i.e. I would need to have the full banner of the generation and/or
> the proc_card, the param_card, the run_card and potentially the model if
> this is not one of those that can be auto-downloaed in 2.6.1
>
>> On 26 Jan 2018, at 23:52, Animesh Chatterjee <email address hidden> wrote:
>>
>> New question #663643 on MadGraph5_aMC@NLO:
>> https://answers.launchpad.net/mg5amcnlo/+question/663643
> Question #663643 : Questions : MadGraph5_aMC@NLO<https://answers.launchpad.net/mg5amcnlo/+question/663643>
> answers.launchpad.net
> I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.
>
>
>>
>> I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.
>>
>> --
>> You received this question notification because you are an answer
>> contact for MadGraph5_aMC@NLO.
>
> --
> If this answers your question, please go to the following page to let us
> know that it is solved:
> https://answers.launchpad.net/mg5amcnlo/+question/663643/+confirm?answer_id=0
>
> If you still need help, you can reply to this email or go to the
> following page to enter your feedback:
> https://answers.launchpad.net/mg5amcnlo/+question/663643
> Question #663643 : Questions : MadGraph5_aMC@NLO<https://answers.launchpad.net/mg5amcnlo/+question/663643>
> answers.launchpad.net
> I was previously using MG5_aMC_v2_2_3 version, but now I have updated it to MG5_aMC_v2_6_1 version. I am getting different results for same parameters. The model I am using is the same, but in the earlier version I was able to generate events with one of the beam energy set to zero, but in the latest one I am getting error on phase space. I am not able to understand the difference.
>
>
>
> You received this question notification because you asked the question.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said :
#4

Hi,

Since Kim, describe to me the problem today, I think that the following patch might solve the issue:

=== modified file 'madgraph/iolibs/template_files/madevent_symmetry.f'
--- madgraph/iolibs/template_files/madevent_symmetry.f 2016-11-16 23:59:06 +0000
+++ madgraph/iolibs/template_files/madevent_symmetry.f 2018-03-16 11:15:27 +0000
@@ -102,6 +102,8 @@
          if (abs(lpp(2)) .eq. 1 .or. abs(lpp(2)) .eq. 2) m2 = 0.938d0
          if (abs(lpp(1)) .eq. 3) m1 = 0.000511d0
          if (abs(lpp(2)) .eq. 3) m2 = 0.000511d0
+ if(ebeam(1).lt.m1) ebeam(1) = m1
+ if(ebeam(2).lt.m2) ebeam(2) = m2
          pi1(0)=ebeam(1)
          pi1(3)=sqrt(max(ebeam(1)**2-m1**2,0d0))
          pi2(0)=ebeam(2)

Not sure if it fixes your problem but this should work for similar issue.
Now for the second problem that Kim explained to me, I really need those additional input to be able to comment.

Cheers,

Olivier

Revision history for this message
Garrett (thegrb93) said :
#5

Thanks so much for the responses Olivier,
I apologize for my unresponsive colleague, Animesh. I've been waiting 3 months for him to find a solution to this and I'm shocked to find this thread he's not even responded to when offered help.

I was unable to match my theta distributions with his MG5_aMC_v2_2_3 ones as well and I'm using latest Madgraph.
The results we are trying to replicate are from https://arxiv.org/abs/1410.1566 in section 5.1. I believe they used the older version of MadGraph and reportedly match his MG5_aMC_v2_2_3 distributions.

I've packaged the model here https://www.dropbox.com/s/sbekuayekalgd1a/DarkPhoDM.tar.gz?raw=1
And an example proc_card for madevent
generate_events DM_3.000000_1.000000_0.010000_0.500000
set ebeam1 0.938272
set ebeam2 120.000000
set vpmass 3.000000
set chimass 1.000000
set mguser 1 0.010000
set mguser 2 0.500000

I would like to try MG5_aMC_v2_2_3 myself, but I need MG5_aMC_v2_2_3. I've searched the madgraph homepage and can't find a download for it. Any idea where I can get it?

Please let me know if you see any issue with the model or reason for it to be discrepant between madgraph versions. Thanks.

Revision history for this message
Garrett (thegrb93) said :
#6

Also, I should mention I've always used a beam energy of 0.938272, so that patch you mentioned has no affect on my results.

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said :
#7

Hi,

I have try in 2.2.3 and this model is simply crashing in that version, while it produces result on the latest version (5.8 pb).

I try to make it work in 2.2.3 but looks like that the model has some issue with the param_card which is not read correctly
and the mass of the chi stays at his default value (100). Obviously I do not really want to debug what is the problem in that version
since I clearly fix the issue in a future version.

Now I have found that your command in 1.5.5 was returning 7.6pb so a quite larger number actually.

But they are one obvious difference between those two runs and one subtil one.
1) the default pdf changed between the two version
2) it is possible that the dynamical scale choice change as well (did not quantify this point)

so if I run the following script in both version:

> import model_v4 DarkPhoDM
> generate p p > chi chi~
> output
> launch
> set pdlabel cteq6l1
> set fixed_ren_scale True
> set fixed_fac_scale True
> set scale 11.188
> set dsqrt_q2fact1 11.188
> set dsqrt_q2fact2 11.188
> set ebeam1 0.938272
> set ebeam2 120.000000
> set vpmass 3.000000
> set chimass 1.000000
> set mguser 1 0.010000
> set mguser 2 0.500000

then in 1.5.5 it gives: 5.601(7)
and in 2.6.1 it gives 5.593 (7)

Which sounds reasonable. I guess that the main effect between the two version is the PDF but I have not check it.

Cheers,

Olivier

> On 11 Apr 2018, at 23:32, Garrett <email address hidden> wrote:
>
> Question #663643 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/663643
>
> Garrett requested more information:
> Thanks so much for the responses Olivier,
> I apologize for my unresponsive colleague, Animesh. I've been waiting 3 months for him to find a solution to this and I'm shocked to find this thread he's not even responded to when offered help.
>
> I was unable to match my theta distributions with his MG5_aMC_v2_2_3 ones as well and I'm using latest Madgraph.
> The results we are trying to replicate are from https://arxiv.org/abs/1410.1566 in section 5.1. I believe they used the older version of MadGraph and reportedly match his MG5_aMC_v2_2_3 distributions.
>
> I've packaged the model here https://www.dropbox.com/s/sbekuayekalgd1a/DarkPhoDM.tar.gz?raw=1
> And an example proc_card for madevent
> generate_events DM_3.000000_1.000000_0.010000_0.500000
> set ebeam1 0.938272
> set ebeam2 120.000000
> set vpmass 3.000000
> set chimass 1.000000
> set mguser 1 0.010000
> set mguser 2 0.500000
>
> I would like to try MG5_aMC_v2_2_3 myself, but I need MG5_aMC_v2_2_3.
> I've searched the madgraph homepage and can't find a download for it.
> Any idea where I can get it?
>
> Please let me know if you see any issue with the model or reason for it
> to be discrepant between madgraph versions. Thanks.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Garrett (thegrb93) said :
#8

Sorry for my long hiatus, I've tested your script and it does yield good agreement with the cross section, but kinematics are still very different.

Here's a dark matter theta distribution between the two version using your script.
https://www.dropbox.com/s/uvzm8unzz2h7tl5/OldNew.png?dl=0

Can you help with this problem?

Provide an answer of your own, or ask Animesh Chatterjee for more information if necessary.

To post a message you must log in.