MadGraph5_aMC@NLO

Incorrect jet cross section for 5f at fNLO

Asked by jannisgebbo

Dear Madgraph experts,

I am generating the process "p p > j j [QCD]" and looking at the inclusive jet spectrum for jets with p_T > 40 GeV at fixed order (so no showering). I am comparing my results with theory calculations obtained with NNLOJet at LO as well as NLO.

Using the default Standard Model with 4 massless quark flavors, LO as well as NLO results are within ~5% agreement. After realizing that NNLOJet results were using a 5f scheme, I regenerated the same process with 5 massless quark flavors via "import model loop_sm-no_b_mass". Now, LO results were in perfect agreement (<1%) while at NLO, however, the inclusive jet cross section was completely off by up to factors of 10. The NLO spectrum is even significantly below the LO spectrum. Hence, the NLO computation seems not to be working correctly.
I am aware that for this process at NLO, one has to implement asymmetric p_t cuts on the jets. I tried setting the pt_cut of jet1 between 5 and 15 GeV higher than the cut for jet2 (e.g. cutting jet1 at 40 GeV and jet2 at 25 GeV). Additionally, I am using PDF sets which are suitable for 5f NLO calculations.

This is extremely puzzling to me, especially because with a 4f scheme this same problem is not occurring and at this point, I am running out of ideas how to fix this. I hope that you can help me finding where this issue may be coming from.

Thank you in advance for your help!

Best,
Jannis

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
marco zaro Edit question
Solved by:
jannisgebbo
Solved:
Last query:
Last reply:
Revision history for this message
marco zaro (marco-zaro) said : 		#1

Dear Jannis,
would you please send me by email the relevant files necessary to reproduce your numbers (proc_card, run_card, param_card, cuts.f, setscales.f, whatever else was changed w.r.t the original ones in the template).
Also, can you please tell me the results you expect at LO and NLO?
Finally, this is for dijet production, not inclusive jet production, is this correct?

Thanks

Marco

Revision history for this message
jannisgebbo (jannisgebbo) said : 		#2

Dear Marco,
thank you very much for dealing with my issue. To this mail, I attached all the files that differ from the template and that are needed to reproduce my run.
Moreover, I attached a little PDF that should briefly summarize what I was trying to do and how you can reproduce my comparison. In there, I tried to describe what I mean with inclusive jet production (not dijet production).
Finally, I also included the differential cross sections (dsigma/p_T dp_T) from NNLOJet computations for NLO as well as LO.

I hope, everything is clear more or less. If not, I will be happy to further explain my computation.

Thanks again! Have a nice weekend and Cheers,
Jannis

Am 21.09.23, 10:10 schrieb "<email address hidden> im Auftrag von marco zaro" <<email address hidden> im Auftrag von <email address hidden>>:

    Your question #707939 on MadGraph5_aMC@NLO changed:
    https://answers.launchpad.net/mg5amcnlo/+question/707939

        Status: Open => Needs information

    marco zaro requested more information:
    Dear Jannis,
    would you please send me by email the relevant files necessary to reproduce your numbers (proc_card, run_card, param_card, cuts.f, setscales.f, whatever else was changed w.r.t the original ones in the template).
    Also, can you please tell me the results you expect at LO and NLO?
    Finally, this is for dijet production, not inclusive jet production, is this correct?

    Thanks

    Marco

    --
    To answer this request for more information, you can either reply to
    this email or enter your reply at the following page:
    https://answers.launchpad.net/mg5amcnlo/+question/707939

    You received this question notification because you asked the question.

Revision history for this message
marco zaro (marco-zaro) said : 		#3

Hi Jannis,
you cannot attach files to launchpad questions, so you need to send them to me via email or any file-share system…
Thanks

Marco

> On 22 Sep 2023, at 16:45, Jannis Gebhard <email address hidden> wrote:
>
> Question #707939 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/707939
>
> Status: Needs information => Open
>
> Jannis Gebhard gave more information on the question:
> Dear Marco,
> thank you very much for dealing with my issue. To this mail, I attached all the files that differ from the template and that are needed to reproduce my run.
> Moreover, I attached a little PDF that should briefly summarize what I was trying to do and how you can reproduce my comparison. In there, I tried to describe what I mean with inclusive jet production (not dijet production).
> Finally, I also included the differential cross sections (dsigma/p_T dp_T) from NNLOJet computations for NLO as well as LO.
>
> I hope, everything is clear more or less. If not, I will be happy to
> further explain my computation.
>
> Thanks again! Have a nice weekend and Cheers,
> Jannis
>
>
> Am 21.09.23, 10:10 schrieb "<email address hidden> im Auftrag von marco
> zaro" <<email address hidden> im Auftrag von
> <email address hidden>>:
>
> Your question #707939 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/707939
>
> Status: Open => Needs information
>
> marco zaro requested more information:
> Dear Jannis,
> would you please send me by email the relevant files necessary to reproduce your numbers (proc_card, run_card, param_card, cuts.f, setscales.f, whatever else was changed w.r.t the original ones in the template).
> Also, can you please tell me the results you expect at LO and NLO?
> Finally, this is for dijet production, not inclusive jet production, is this correct?
>
> Thanks
>
> Marco
>
> --
> To answer this request for more information, you can either reply to
> this email or enter your reply at the following page:
> https://answers.launchpad.net/mg5amcnlo/+question/707939
>
> You received this question notification because you asked the
> question.
>
> --
> You received this question notification because you are assigned to this
> question.

Revision history for this message
jannisgebbo (jannisgebbo) said : 		#4

Update:

I managed to resolve this issue and get the correct fixed order cross section by noticing that the value of the 'ickkw' parameter was set to 4 (UNLOPS merging) instead of 0 (no merging). This produced additional terms which are used if the fixed order cross section is merged to a parton shower. In the case of fixed-order calculations, this messed up the cross section.

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

Thanks a lot for this update,

I will put in place a check that such flag is set to zero in such setup (or make the code to crash) to try to avoid such type of issue in the future.

Thanks,

Olivier

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#6

A crash has been added (will be release in 3.5.3)

Thanks,

Olivier