MadGraph5_aMC@NLO

Madweight code documentation

Asked by Anne-Catherine Le Bihan

Hi,

I'm trying to run the MEM for processes like :

import model sm-ckm
define l+ = e+ mu+
define l- = e- mu-
define v+ = ve~ vm~
define v- = ve vm
generate p p > t t~ h, h > w+ w- > l+ v- l- v+, (t > b w+, w+ > l+ v-), (t~ > b~ w-, w- > j j)
add process p p > t t~ h, h > w+ w- > l+ v- l- v+, (t > b w+, w+ > j j), (t~ > b~ w-, w- > l- v+)
add process p p > t t~ h, h > w+ w- > j j l- v+, (t > b w+, w+ > l+ v-), (t~ > b~ w-, w- > l- v+)
add process p p > t t~ h, h > w+ w- > l+ v- j j, (t > b w+, w+ > l+ v-), (t~ > b~ w-, w- > l- v+)

Trying to run on generated ttH, H-> WW in 3 leptons events, I get an output (/Events/fermi/weights.out) with value and integration_error : 0.0 0.0

Looking at the code I see that the jacobian values are somehow diverging (cf l 467-470 of SubProcesses/gen_ps.f) :

JAC 3.6073923971296544E-002
JAC 32.578022390710395
JAC 19136.445405684895
JAC 47962.413019255488
JAC 193853724.39522931
JAC 5502006171513.9121
JAC 2828362007011530.0

How can I further debug this or find documentation about the code organisation and cross-checks to do ?

Thanks,

Anne-Catherine

Question information

Language:
English Edit question
Status:
Needs information
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Pierre Artoisenet Edit question
Last query:
Last reply:
Revision history for this message
Pierre Artoisenet (partois) said : 		#1

Hi Anne Catherine,

Are you using the lastest version of madgraph ?
(MG5_aMC_v2_1_1)

In which routine do you print the value of the jacobian ?
(line 469 is

Breit_Wigner_for_part(MG_num,M,W)

in gen_ps.f)

Can you also give me the definiton of your transfer function,
so that I can reproduce your problem ?

Pierre

Revision history for this message
Anne-Catherine Le Bihan (anne-catherine-le-bihan) said : 		#2

Hi Pierre,

Sorry for the late reply.

I have reproduced the problem with MG5_aMC_v2_1_1.

The value of the jacobian is printed after l 460 ( subroutine generate_propa(x,n_var)).

I'm using the default transfer functions (change_tf 1, double gaussians for jets) on ttH events generated with madgraph using the same setup. (The lhe file has been converted into an lhco file.)

Thanks for your help,

Anne-Catherine

Revision history for this message
Kirill Skovpen (kirill-skovpen) said : 		#3

Hi Pierre,

Sorry for jumping in. Just to add that we would highly appreciate any help on the issue Anne-Catherine has described above.

Thanks a lot,

Kirill

Revision history for this message
Kirill Skovpen (kirill-skovpen) said : 		#4

Hello,

Sorry for spamming, but are there any news on this issue ?

Many thanks,

Kirill

Revision history for this message
Pierre Artoisenet (partois) said : 		#5

Hi Kirill,

Sorry for the delay, I will have a look this week

Cheers,
Pierre

Revision history for this message
Kirill Skovpen (kirill-skovpen) said : 		#6

Hello,

Sorry for spamming too much, just to check if there is any progress on this issue ?

Many thanks,

Kirill

Revision history for this message
Pierre Artoisenet (partois) said : 		#7

Hi Kirill,

I am still working on it. The definition of phase-space generator has to be completely rewritten,
which is still ongoing. In any case, this is quite a challenging case for MadWeight. But we should be
fixed by tomorrow.

Pierre

Revision history for this message
Pierre Artoisenet (partois) said : 		#8

A patch has been sent to the authors of this report

Pierre

Revision history for this message
Kirill Skovpen (kirill-skovpen) said : 		#9

Hello Pierre,

Many thanks for the patch.

We are also considering other channels for ttH in addition to 3l channel mentioned above, e.g. 4l:

generate p p > t t~ h, h > e+ ve e- ve~ / z, (t > b w+ , w+ > e+ ve), (t~ > b~ w- , w- > e- ve~)

and also channels where H decays to ZZ and tautau ... Would it be possible to provide patches for all these channels (2l,3l,4l) as well ? Is there some time estimate for such kind of fix(es) if this is possible in general ?

Thanks a lot,

Kirill

Can you help with this problem?

Provide an answer of your own, or ask Anne-Catherine Le Bihan for more information if necessary.

To post a message you must log in.