Check that new diagrams are being used?

Hi,

If you see the diagram then you can be sure that the diagram is indeed used for the computation.
Now the contribution of that diagram can be zero if one of the coupling vanishes/…

If your new diagram is not a four point of interaction it is likely to be linked to on channel of integration.
and we report the contribution of that channel.
Please see here the details on how to find how channel of integration are related to Feynman Diagram:
https://answers.launchpad.net/mg5amcnlo/+question/187387

Cheers,

Olivier

PS: Just for curiosity why do you use the “add model” functionality and not the directly create the full model from FR.
I would say that using FR is safer and more tested than this add model functionality.

On 18 Aug 2015, at 01:12, Matthew Lockner <email address hidden> wrote:

> New question #270481 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Hello,
>
> I'm trying to use MadGraph to reproduce a result from this (http://iopscience.iop.org/0954-3899/29/3/308). Please pardon my referencing specifics from there - it would be difficult to try to write equations readably with this interface.
>
> The paper defines a piece of Lagrangian (Eq. 4) involving two parameters c_t and d_t (Eq. 6). This yields new Feynman rules and additional diagrams to those given by the standard model (Figs. 1 and 2). Then for each event we can compute a value of an observable, here called f_2 (Eq. 10) whose distribution is similar to a Gaussian (Figure 7). As suggested by, e.g., Table 4, the mean of the f_2 distribution should increase as the value of the parameter d_t is increased.
>
> Using FeynRules, I wrote the piece of Lagrangian in Eq. 4 and output the resulting new rules using WriteUFO. I added the directory output to the models/ subdirectory of madgraph5. I then start mg5_aMC and run "add model Standard_Model_UFO --recreate" (I accept blame for the poor choice of name) which seems to succeed without errors. Then I run
>
> MG5_aMC>generate g g > t t~ > b b~ mu+ vm mu- vm~ QED=4 QCD=3
>
> This also succeeds, and I use "display diagrams" to check the diagrams - and indeed I see the new seagull diagram from Fig. 2, which is encouraging. I then run output and launch. I choose to edit param_card.dat, which does indeed appear to have my new parameters in it:
>
> ###################################
> ## INFORMATION FOR FRBLOCK
> ###################################
> Block frblock
> 1 2.500000e-01 # ct
> 2 4.000000e-01 # dt
>
> At this point I can generate a set of events for whichever values of ct and dt I wish.
>
> The problem comes when I try to plot the distribution of f_2 for different values of ct and dt. As best as I can tell, the means of the gaussians are not changing with any kind of dependence on ct or dt (as evidenced when I plotted the means for varying values of both).
>
> I need to check that Madgraph is actually ever using the new Feynman diagrams or perhaps find out for a given event which diagram was selected. Can you think of anything within MadGraph that might be helpful?
>
> Thanks,
> Matthew Lockner
>
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

On further investigation, the event sets appear to come out identical (at least on reading through the first dozen or so events from each unweighted_events.lhe file) regardless of my settings of ct and dt. The new diagram (seagull diagram) is indeed a four-participant ggtt~ diagram.

I used "add model" just because it made intuitive sense as the command I'd need. Your suggestion instead would be to merge the models within Feynrules? Is the standard model defined by Feynrules the same as that used by default in MadGraph?

Thanks,
Matthew Lockner

Hi,

> The
> new diagram (seagull diagram) is indeed a four-participant ggtt~
> diagram.

ok, then you do not have a specific channel of integration related to that Feynman diagram.

> On further investigation, the event sets appear to come out identical
> (at least on reading through the first dozen or so events from each
> unweighted_events.lhe file) regardless of my settings of ct and dt.

The reason that you do not see anything is likely that… they are nothing to see meaning that your contribution is very small compare to the SM one.

Did you try extremely large value of ct/dt? just to check that you indeed have an effect when you have large coupling.
Did you try to single out that diagram? to see what is the contribution of the amplitude square for the value that you consider?
(you can do that if you define a new order —in top of QED/QCD— for your seagull coupling)

> I used "add model" just because it made intuitive sense as the command
> I'd need. Your suggestion instead would be to merge the models within
> Feynrules?

Yes.

> Is the standard model defined by Feynrules the same as that
> used by default in MadGraph?

The one of FR is more general (CKM/ mass for the light quark/…) but ours is a simple restriction of the FeynRules one.
If you want to have the restriction file used to produce the SM of MG, you can ask Claude Duhr for it.
(in top of that we have add a couple of restrict_XXX.dat in the UFO file in order to apply additional restriction but at the MG level)

Cheers,

Olivier

On 19 Aug 2015, at 01:57, Matthew Lockner <email address hidden> wrote:

> Question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Status: Answered => Open
>
> Matthew Lockner is still having a problem:
> On further investigation, the event sets appear to come out identical
> (at least on reading through the first dozen or so events from each
> unweighted_events.lhe file) regardless of my settings of ct and dt. The
> new diagram (seagull diagram) is indeed a four-participant ggtt~
> diagram.
>
> I used "add model" just because it made intuitive sense as the command
> I'd need. Your suggestion instead would be to merge the models within
> Feynrules? Is the standard model defined by Feynrules the same as that
> used by default in MadGraph?
>
> Thanks,
> Matthew Lockner
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

On Wed, Aug 19, 2015 at 2:57 AM, Olivier Mattelaer <
<email address hidden>> wrote:

>
> Did you try extremely large value of ct/dt? just to check that you indeed
> have an effect when you have large coupling.
>

I tried ct=0 with dt=2000 and dt=2,000,000, and no apparent difference in
my plot. I am using unweighted_events.lhe - does that seem like the
correct file to look in?

> Did you try to single out that diagram? to see what is the contribution of
> the amplitude square for the value that you consider?
> (you can do that if you define a new order —in top of QED/QCD— for your
> seagull coupling)
>

Can you elaborate on how to do this?

(At this point I am also using the model merged within Feynrules and not
"add model"; my piece of Lagrangian is called L5 and the SM Lagrangian is
LSM, so I called WriteUFO[LSM + L5], and then the "import model" command
within FeynRules).

Thanks,
Matthew Lockner

Olivier,

Firstly thank you for all your help so far. Could you give me more
information or steps I need to take for the following suggestion?:

> Did you try to single out that diagram? to see what is the contribution of
> the amplitude square for the value that you consider?
> (you can do that if you define a new order —in top of QED/QCD— for your
> seagull coupling)
>

Thanks,
Matthew Lockner

On Mon, Aug 24, 2015 at 6:36 PM, Matthew Lockner <
<email address hidden>> wrote:

> Your question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Status: Answered => Open
>
> You are still having a problem:
> On Wed, Aug 19, 2015 at 2:57 AM, Olivier Mattelaer <
> <email address hidden>> wrote:
>
> >
> > Did you try extremely large value of ct/dt? just to check that you indeed
> > have an effect when you have large coupling.
> >
>
> I tried ct=0 with dt=2000 and dt=2,000,000, and no apparent difference in
> my plot. I am using unweighted_events.lhe - does that seem like the
> correct file to look in?
>
>
> > Did you try to single out that diagram? to see what is the contribution
> of
> > the amplitude square for the value that you consider?
> > (you can do that if you define a new order —in top of QED/QCD— for your
> > seagull coupling)
> >
>
> Can you elaborate on how to do this?
>
> (At this point I am also using the model merged within Feynrules and not
> "add model"; my piece of Lagrangian is called L5 and the SM Lagrangian is
> LSM, so I called WriteUFO[LSM + L5], and then the "import model" command
> within FeynRules).
>
> Thanks,
> Matthew Lockner
>
> --
> You received this question notification because you asked the question.
>

Hi,

You have to define an additional entry in the coupling_order.py
and then track the coupling that you use and replace the associate coupling order (which will be QCD and/or QED to the new one define)

Cheers,

Olivier

On 31 Aug 2015, at 19:51, Matthew Lockner <email address hidden> wrote:

> Question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Matthew Lockner gave more information on the question:
> Olivier,
>
> Firstly thank you for all your help so far. Could you give me more
> information or steps I need to take for the following suggestion?:
>
>> Did you try to single out that diagram? to see what is the contribution of
>> the amplitude square for the value that you consider?
>> (you can do that if you define a new order —in top of QED/QCD— for your
>> seagull coupling)
>>
>
> Thanks,
> Matthew Lockner
>
>
> On Mon, Aug 24, 2015 at 6:36 PM, Matthew Lockner <
> <email address hidden>> wrote:
>
>> Your question #270481 on MadGraph5_aMC@NLO changed:
>> https://answers.launchpad.net/mg5amcnlo/+question/270481
>>
>> Status: Answered => Open
>>
>> You are still having a problem:
>> On Wed, Aug 19, 2015 at 2:57 AM, Olivier Mattelaer <
>> <email address hidden>> wrote:
>>
>>>
>>> Did you try extremely large value of ct/dt? just to check that you indeed
>>> have an effect when you have large coupling.
>>>
>>
>> I tried ct=0 with dt=2000 and dt=2,000,000, and no apparent difference in
>> my plot. I am using unweighted_events.lhe - does that seem like the
>> correct file to look in?
>>
>>
>>> Did you try to single out that diagram? to see what is the contribution
>> of
>>> the amplitude square for the value that you consider?
>>> (you can do that if you define a new order —in top of QED/QCD— for your
>>> seagull coupling)
>>>
>>
>> Can you elaborate on how to do this?
>>
>> (At this point I am also using the model merged within Feynrules and not
>> "add model"; my piece of Lagrangian is called L5 and the SM Lagrangian is
>> LSM, so I called WriteUFO[LSM + L5], and then the "import model" command
>> within FeynRules).
>>
>> Thanks,
>> Matthew Lockner
>>
>> --
>> 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.

I see the coupling_orders.py for my model; it's very short and now only has
QCD and QED. But I am afraid I don't quite understand the remainder of
your response. How would I associate that order with (for instance) my
seagull diagram, and how does that help me find its contribution to
amplitude squared? My apologies, I am still quite new to MadGraph and not
familiar with its internals.

On Mon, Aug 31, 2015 at 5:46 PM, Olivier Mattelaer <
<email address hidden>> wrote:

> Your question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Status: Open => Answered
>
> Olivier Mattelaer proposed the following answer:
> Hi,
>
> You have to define an additional entry in the coupling_order.py
> and then track the coupling that you use and replace the associate
> coupling order (which will be QCD and/or QED to the new one define)
>
> Cheers,
>
> Olivier
>
> On 31 Aug 2015, at 19:51, Matthew Lockner
> <email address hidden> wrote:
>
> > Question #270481 on MadGraph5_aMC@NLO changed:
> > https://answers.launchpad.net/mg5amcnlo/+question/270481
> >
> > Matthew Lockner gave more information on the question:
> > Olivier,
> >
> > Firstly thank you for all your help so far. Could you give me more
> > information or steps I need to take for the following suggestion?:
> >
> >> Did you try to single out that diagram? to see what is the contribution
> of
> >> the amplitude square for the value that you consider?
> >> (you can do that if you define a new order —in top of QED/QCD— for your
> >> seagull coupling)
> >>
> >
> > Thanks,
> > Matthew Lockner
> >
> >
> > On Mon, Aug 24, 2015 at 6:36 PM, Matthew Lockner <
> > <email address hidden>> wrote:
> >
> >> Your question #270481 on MadGraph5_aMC@NLO changed:
> >> https://answers.launchpad.net/mg5amcnlo/+question/270481
> >>
> >> Status: Answered => Open
> >>
> >> You are still having a problem:
> >> On Wed, Aug 19, 2015 at 2:57 AM, Olivier Mattelaer <
> >> <email address hidden>> wrote:
> >>
> >>>
> >>> Did you try extremely large value of ct/dt? just to check that you
> indeed
> >>> have an effect when you have large coupling.
> >>>
> >>
> >> I tried ct=0 with dt=2000 and dt=2,000,000, and no apparent difference
> in
> >> my plot. I am using unweighted_events.lhe - does that seem like the
> >> correct file to look in?
> >>
> >>
> >>> Did you try to single out that diagram? to see what is the contribution
> >> of
> >>> the amplitude square for the value that you consider?
> >>> (you can do that if you define a new order —in top of QED/QCD— for your
> >>> seagull coupling)
> >>>
> >>
> >> Can you elaborate on how to do this?
> >>
> >> (At this point I am also using the model merged within Feynrules and not
> >> "add model"; my piece of Lagrangian is called L5 and the SM Lagrangian
> is
> >> LSM, so I called WriteUFO[LSM + L5], and then the "import model" command
> >> within FeynRules).
> >>
> >> Thanks,
> >> Matthew Lockner
> >>
> >> --
> >> 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.
>
> --
> 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/270481/+confirm?answer_id=5
>
> 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/270481
>
> You received this question notification because you asked the question.
>

Hi,

a diagram is composed by a suite of vertex (each vertex is define in vertex.py)
each vertex is associate to one or more coupling (this is part of the vertex definition)
each of those couplings (which are define in couplings.py) are associate to a coupling order.

The coupling order can be use to restrict the type of diagram include in the computation
for example
generate p p > t t~ QCD=0
is not the same as
generate p p > t t~ QED=0

More details on the meaning of those syntax is available here:
https://cp3.irmp.ucl.ac.be/projects/madgraph/attachment/wiki/MGTalks/13_06_10_tutomg_tasi.pdf

So the idea is to define a new order in order to single out the contribution that you are looking at.

Cheers,

Olivier

On 01 Sep 2015, at 22:16, Matthew Lockner <email address hidden> wrote:

> Question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Status: Answered => Open
>
> Matthew Lockner is still having a problem:
> I see the coupling_orders.py for my model; it's very short and now only has
> QCD and QED. But I am afraid I don't quite understand the remainder of
> your response. How would I associate that order with (for instance) my
> seagull diagram, and how does that help me find its contribution to
> amplitude squared? My apologies, I am still quite new to MadGraph and not
> familiar with its internals.
>
> On Mon, Aug 31, 2015 at 5:46 PM, Olivier Mattelaer <
> <email address hidden>> wrote:
>
>> Your question #270481 on MadGraph5_aMC@NLO changed:
>> https://answers.launchpad.net/mg5amcnlo/+question/270481
>>
>> Status: Open => Answered
>>
>> Olivier Mattelaer proposed the following answer:
>> Hi,
>>
>> You have to define an additional entry in the coupling_order.py
>> and then track the coupling that you use and replace the associate
>> coupling order (which will be QCD and/or QED to the new one define)
>>
>> Cheers,
>>
>> Olivier
>>
>> On 31 Aug 2015, at 19:51, Matthew Lockner
>> <email address hidden> wrote:
>>
>>> Question #270481 on MadGraph5_aMC@NLO changed:
>>> https://answers.launchpad.net/mg5amcnlo/+question/270481
>>>
>>> Matthew Lockner gave more information on the question:
>>> Olivier,
>>>
>>> Firstly thank you for all your help so far. Could you give me more
>>> information or steps I need to take for the following suggestion?:
>>>
>>>> Did you try to single out that diagram? to see what is the contribution
>> of
>>>> the amplitude square for the value that you consider?
>>>> (you can do that if you define a new order —in top of QED/QCD— for your
>>>> seagull coupling)
>>>>
>>>
>>> Thanks,
>>> Matthew Lockner
>>>
>>>
>>> On Mon, Aug 24, 2015 at 6:36 PM, Matthew Lockner <
>>> <email address hidden>> wrote:
>>>
>>>> Your question #270481 on MadGraph5_aMC@NLO changed:
>>>> https://answers.launchpad.net/mg5amcnlo/+question/270481
>>>>
>>>> Status: Answered => Open
>>>>
>>>> You are still having a problem:
>>>> On Wed, Aug 19, 2015 at 2:57 AM, Olivier Mattelaer <
>>>> <email address hidden>> wrote:
>>>>
>>>>>
>>>>> Did you try extremely large value of ct/dt? just to check that you
>> indeed
>>>>> have an effect when you have large coupling.
>>>>>
>>>>
>>>> I tried ct=0 with dt=2000 and dt=2,000,000, and no apparent difference
>> in
>>>> my plot. I am using unweighted_events.lhe - does that seem like the
>>>> correct file to look in?
>>>>
>>>>
>>>>> Did you try to single out that diagram? to see what is the contribution
>>>> of
>>>>> the amplitude square for the value that you consider?
>>>>> (you can do that if you define a new order —in top of QED/QCD— for your
>>>>> seagull coupling)
>>>>>
>>>>
>>>> Can you elaborate on how to do this?
>>>>
>>>> (At this point I am also using the model merged within Feynrules and not
>>>> "add model"; my piece of Lagrangian is called L5 and the SM Lagrangian
>> is
>>>> LSM, so I called WriteUFO[LSM + L5], and then the "import model" command
>>>> within FeynRules).
>>>>
>>>> Thanks,
>>>> Matthew Lockner
>>>>
>>>> --
>>>> 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.
>>
>> --
>> 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/270481/+confirm?answer_id=5
>>
>> 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/270481
>>
>> 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.

OK so I added a new coupling MYCOUPLING to coupling_orders.py:

MYCOUPLING = CouplingOrder(name = 'MYCOUPLING',
                           expansion_order=99,
                           hierarchy=2)

and modified couplings.py to include copies of the original coupling entries, but for MYCOUPLING:

GC_137 = Coupling(name = 'GC_137',
                 value = '(ct*G**2)/MT',
                 order = {'MYCOUPLING':1})

GC_138 = Coupling(name = 'GC_138',
                 value = '(dt*complex(0,1)*G**2)/MT',
                 order = {'MYCOUPLING':1})

and finally modified vertices.py to specify use of my new coupling:

V_164 = Vertex(name = 'V_164',
               particles = [ P.t__tilde__, P.t, P.g, P.g ],
               color = [ 'f(-1,3,4)*T(-1,2,1)' ],
               lorentz = [ L.FFVV1, L.FFVV2 ],
               couplings = {(0,0):C.GC_137,(0,1):C.GC_138})

Now if I do "generate g g > t t~ MYCOUPLING=1" I do see the seagull diagram, but it included all the others as having MYCOUPLING=0 though I never specified a MYCOUPLING for those other ten vertices. Can I make it only include the MYCOUPLING=1 diagram and not the MYCOUPLING=0, and how?

Thanks,
Matthew Lockner

> Now if I do "generate g g > t t~ MYCOUPLING=1" I do see the seagull diagram, but it included all the others as having MYCOUPLING=0 though I never specified a MYCOUPLING for those other ten vertices. Can I make it only include the MYCOUPLING=1 diagram and not the MYCOUPLING=0, and how?

That’s why I refer you to the tutorial. Since you need to understand how those coupling works.

Cheers,

Olivier
On 02 Sep 2015, at 00:56, Matthew Lockner <email address hidden> wrote:

> Question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Status: Answered => Open
>
> Matthew Lockner is still having a problem:
> OK so I added a new coupling MYCOUPLING to coupling_orders.py:
>
> MYCOUPLING = CouplingOrder(name = 'MYCOUPLING',
> expansion_order=99,
> hierarchy=2)
>
> and modified couplings.py to include copies of the original coupling
> entries, but for MYCOUPLING:
>
> GC_137 = Coupling(name = 'GC_137',
> value = '(ct*G**2)/MT',
> order = {'MYCOUPLING':1})
>
> GC_138 = Coupling(name = 'GC_138',
> value = '(dt*complex(0,1)*G**2)/MT',
> order = {'MYCOUPLING':1})
>
>
> and finally modified vertices.py to specify use of my new coupling:
>
> V_164 = Vertex(name = 'V_164',
> particles = [ P.t__tilde__, P.t, P.g, P.g ],
> color = [ 'f(-1,3,4)*T(-1,2,1)' ],
> lorentz = [ L.FFVV1, L.FFVV2 ],
> couplings = {(0,0):C.GC_137,(0,1):C.GC_138})
>
>
> Now if I do "generate g g > t t~ MYCOUPLING=1" I do see the seagull diagram, but it included all the others as having MYCOUPLING=0 though I never specified a MYCOUPLING for those other ten vertices. Can I make it only include the MYCOUPLING=1 diagram and not the MYCOUPLING=0, and how?
>
> Thanks,
> Matthew Lockner
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Well as I'm understanding it, giving QED=n means use vertices that have QED coupling of n or lower, likewise QCD=m includes QCD vertices of coupling order m or lower. My guess is that to have it exclude all but one diagram, I'd actually need the vertices I wish included to have MYCOUPLING=0 and all other participant vertices to have MYCOUPLING=1; then I should be able to exclude all but the desired vertex by including MYCOUPLING=0. However, that would involve making many modifications to my vertices.py, a likely error-prone process, so I wanted to check first if there isn't a more efficient way.

Also I presume the order argument in couplings.py entries is just a Python dict, correct?

HI,

I was thinking that doing something like
QED=0 QCD=0
or
QED=1 QCD=0
or
QED=0 QCD=0
should do it. Is this not working?

Cheers,

Olivier
On 02 Sep 2015, at 02:12, Matthew Lockner <email address hidden> wrote:

> Question #270481 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/270481
>
> Status: Answered => Open
>
> Matthew Lockner is still having a problem:
> Well as I'm understanding it, giving QED=n means use vertices that have
> QED coupling of n or lower, likewise QCD=m includes QCD vertices of
> coupling order m or lower. My guess is that to have it exclude all but
> one diagram, I'd actually need the vertices I wish included to have
> MYCOUPLING=0 and all other participant vertices to have MYCOUPLING=1;
> then I should be able to exclude all but the desired vertex by including
> MYCOUPLING=0. However, that would involve making many modifications to
> my vertices.py, a likely error-prone process, so I wanted to check first
> if there isn't a more efficient way.
>
> Also I presume the order argument in couplings.py entries is just a
> Python dict, correct?
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.