MadGraph5_aMC@NLO

process h > g g g surprisingly inefficient

Asked by Stephen Mrenna

I am studying the process h > g g g using:
import model heft
generate h > g g g HIG=1
I use:
1= ckkw
10.0 = xqcut
0 = lpp1
0 = lpp2 (don't know if these matter)
125.0 = scale

With a target of 10k events, I am getting typically 70-80 at the end. This doesn't sound right, and is discouraging, since I am aiming for 100k or so, plus want to also add in h > g g g g.
I don't find the treatment of decays as well documented as for processes (but maybe this is my ignorance).
Any suggestions on what I might do to increase the efficiency?

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
Stephen Mrenna (stephen-mrenna) said : 		#1

As a followup, I bypassed the ckkw/xqcut flags (set them to 0.0) and set only an mjj cut.
This was lightning fast and very efficient.
I assume that all the powers of alpha_s are evaluated at 125 GeV, which is not ideal.
So, the real question is this: can I generate this process with a kT-cut (or custom cut), and also get the alpha_s reweighting at the vertex, AND maintain the high efficiency.

Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#2

Hi,

I will need to look at the MLM support code to see how much it can handle 1>N.
In general final state radiation (from Z/Higgs/...) are not part of the MLM matching/merging mechanism.
So I have some doubt (and your result seems to confirm that) that the code is supported such kind of MLM process.

Being able to handle this case would make a lot of sense but I need to look at the code to see how much this will be easy/difficult to do. (I hope to be able to do that on Thursday)

Cheers,

Olivier

> On 9 Jul 2019, at 16:57, Stephen Mrenna <email address hidden> wrote:
>
> Question #681881 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/681881
>
> Stephen Mrenna gave more information on the question:
> As a followup, I bypassed the ckkw/xqcut flags (set them to 0.0) and set only an mjj cut.
> This was lightning fast and very efficient.
> I assume that all the powers of alpha_s are evaluated at 125 GeV, which is not ideal.
> So, the real question is this: can I generate this process with a kT-cut (or custom cut), and also get the alpha_s reweighting at the vertex, AND maintain the high efficiency.
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Revision history for this message
Stephen Mrenna (stephen-mrenna) said : 		#3

Thank you!
As far as I can tell, all the interesting stuff is happening in reweight.f.
Best,
Steve

Can you help with this problem?

Provide an answer of your own, or ask Stephen Mrenna for more information if necessary.

To post a message you must log in.