Inquiry on matching+merging in VBF and GGF higgs production

Asked by Iftah Galon on 2019-07-18

Dear MG5 team

Please excuse my confusion in advance.

I am interested in studying Higgs production in association with 2 or more jets.
I am simulating both VBF and GGF processes (separately).

In order to avoid redundancies, I have defined a new UFO model in FeynRules.
I added to the full SM model the effective Higgs-gluon and Higgs-photon couplings, and assigned a new and individual coupling order to each. With this, I am able to control the individual VBF and GGF contributions with respect to QED and QCD.
In addition, I modified the coupling order of the Yukawas to a new coupling order so that I could avoid including diagrams with Yukawa couplings which are negligible.

For each VBF and GGF process, I have generated a (LO) sample and a (LO+1j) sample (= LO + "LO+1j").
For the +1j samples, I turned on MLM matching (and basically use the automated run_card)

The two GGF samples and the LO VBF one ran ok, in both the MG5, and pythia8 stages.
The VBF + 1j sample crashed in the pythia8 stage, with an error message 139 and segmentation fault:
"launch ends with non zero status"

I am wondering whether this is due to a wrong assignment of matching in the VBF case.
I seem to remember reading in past documentation on MG5_launchpad that matching is implemented from the leading order QCD process. In particular, could it be that the matching algorithms fail because VBF (+0j) is a QCD=0 process ?

This sounds a bit weird to me, because I imagine that an initial/final state quark could still have ISR/FSR contributions in the parton shower, regardless of whether its interaction is QED/QCD

In this case, what would be the correct production process.
Is it simply
generate p p > j j h QED=3 QCD=0
add process p p > j j j h QED=3 QCD=1

but with matching turned off ?
Should I reinstate drjj in this case ?

Thank you for your help
Sincerely
Iftah

Question information

Language:
English Edit question
Status:
Open
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Last query:
2019-08-15
Last reply:
2019-08-14
Iftah Galon (iftah) said : #1

One other thing which could be the cause.
I have been letting MG5 guess the value for JetMatching:nJetMax .
For the VBF+1j case it guessed it as 1, and for the GGF case as 2
When I set it to 3 pythia8 works without crashing, but I suspect that this might not be the right way to go

Thank you
Sincerely
Iftah

error 139 is a segfault, you should contact Stefan prestel concerning such segfault.

Could you specify exactly the MG5aMC syntax that you use in each case, this will clarify what you are doing.
I'm especially confused about what you call 0j for GGF

For the VBF, it also depend of your syntax, I would in this case consider the following setup:
JetMatching:nJetMin=2
JetMatching:nJetMax=3
can make a lot of sense to me.

Cheers,

Olivier

Iftah Galon (iftah) said : #3

Hi Olivier,

Thank you for the swift reply.

Indeed it is a segfault which I think is arising due to pythia failing to produce a matched sample with nJetMax =1.

The syntax I use is

for GFF

generate p p > h j j HIGQCD=2 HIGQED=0 YUK=0 QCD=2 QED=0, h > a a
add process p p > h j j HIGQCD=2 HIGQED=0 YUK=0 QCD=3 QED=0, h > a a

and for VBF

generate p p > h j j YUK=0 HIGQCD=0 HIGQED=0 QED=3, h > a a
add process p p > h j j j YUK=0 HIGQCD=0 HIGQED=0 QED=3 QCD=1, h > a a

where the HIGQCD, HIGQED are coupling orders for the effective higgs couplings, and YUK is a coupling order for the SM Yukawas.

Note that I am only generating GGF with njet >= 2

I was reading through

http://home.thep.lu.se/~torbjorn/pythia81html/JetMatching.html

would setting

JetMatching:nJetMin=2 (I did not see a nJetMin in the options - is this nJet ? )
JetMatching:nJetMax=3

imply that the basic incoming hard-process has 2 hard jets, and that the shower should match an extra jet to the added process, i.e. the one coming from the syntax:
add process p p > (hard process) j

Asked differently: Is the number of matched jets <= JetMatching:nJetMax - JetMatching:nJet

Would this also be the correct syntax for the GGF + 2j process ?

Thank you
Sincerely
Iftah

Launchpad Janitor (janitor) said : #4

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

Iftah Galon (iftah) said : #5

Hi Olivier,

Do you think the setting from the last inquiry should work ?

Thank you

Hi,

generate p p > h j j HIGQCD=2 HIGQED=0 YUK=0 QCD=2 QED=0, h > a a
add process p p > h j j HIGQCD=2 HIGQED=0 YUK=0 QCD=3 QED=0, h > a a

This is not a correct syntax for matching/merging since you have the same multiplicity of jet between the two sample (in top of that since "=" means "<=" in MG5aMC, the second syntax includes the first syntax and you have a 100% double counting of the first line.

For VBF:

generate p p > h j j YUK=0 HIGQCD=0 HIGQED=0 QED=3, h > a a
add process p p > h j j j YUK=0 HIGQCD=0 HIGQED=0 QED=3 QCD=1, h > a a

This next comment depend actually of the model, but you typically need QCD=0 in the first line.
(otherwise the default is 99 and you will keep pure QCD diagram that (I guess you do not want(

 JetMatching:nJet =2

seems indeed appropriate for VBF process if you set QCD=0 for the first line (or if your model is written in such way that this is not necessary off course).

Cheers,

Olivier

PS: Obviously you need to validate your plots and in particular the DJR ones.

On 13 Aug 2019, at 17:22, Iftah Galon <<email address hidden><mailto:<email address hidden>>> wrote:

Question #682129 on MadGraph5_aMC@NLO changed:
https://answers.launchpad.net/mg5amcnlo/+question/682129

   Status: Expired => Open

Iftah Galon is still having a problem:
Hi Olivier,

Do you think the setting from the last inquiry should work ?

Thank you

--
You received this question notification because you are an answer
contact for MadGraph5_aMC@NLO.

Iftah Galon (iftah) said : #7

Hi Olivier,

Thank you for the clarifications.
The missing 'j' in GGF syntax was a typo .... thank you for catching that !!

As for the pythia8 card parameters.
I am using MLM, so the parameter to set in the pythia8_card is
JetMatching:nJetMax =

Can one set JetMatching:nJet in the pythia8_card as well, i.e. can I simply add a line
JetMatching:nJet = #value

From
http://home.thep.lu.se/~torbjorn/pythia81html/JetMatching.html

this parameter sets the minimum number of light jets in the incoming process, so setting
JetMatching:nJet = 2
does make sense to me.

I am still not quite sure as to what setting I should give JetMatching:nJetMax .
Is this 3 because there is only one additional hard jet in the hard-process level (from 'add process ...'), or should I increase it beyond 3 ?

Thank you very much
Sincerely
Iftah

Yes you can add such line.

For the max setting it to 3 makes sense as well.

Cheers,

Olivier

> On 14 Aug 2019, at 17:52, Iftah Galon <email address hidden> wrote:
>
> Question #682129 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/682129
>
> Status: Answered => Open
>
> Iftah Galon is still having a problem:
> Hi Olivier,
>
> Thank you for the clarifications.
> The missing 'j' in GGF syntax was a typo .... thank you for catching that !!
>
> As for the pythia8 card parameters.
> I am using MLM, so the parameter to set in the pythia8_card is
> JetMatching:nJetMax =
>
> Can one set JetMatching:nJet in the pythia8_card as well, i.e. can I simply add a line
> JetMatching:nJet = #value
>
>> From
> http://home.thep.lu.se/~torbjorn/pythia81html/JetMatching.html
>
> this parameter sets the minimum number of light jets in the incoming process, so setting
> JetMatching:nJet = 2
> does make sense to me.
>
> I am still not quite sure as to what setting I should give JetMatching:nJetMax .
> Is this 3 because there is only one additional hard jet in the hard-process level (from 'add process ...'), or should I increase it beyond 3 ?
>
>
> Thank you very much
> Sincerely
> Iftah
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Iftah Galon (iftah) said : #9

Thank you Olivier,

I generated a few samples of GGF of this with different xqcut and qcut scales.
I tried to smooth it as possible and for the GGF chose xqcut = 30 GeV and qcut = 50 GeV.

In the resulting DJR plots, should I even be concerned about the 1->0 and 2->1 plots, or should I start only with the higher multiplicity ones (i.e. because the LO is a 2-jet process)

Sincerely
Iftah

Can you help with this problem?

Provide an answer of your own, or ask Iftah Galon for more information if necessary.

To post a message you must log in.