matching problem of bb+1j

Dear Yong,

It looks like that your process is a mix between a 4 flavor and a 5 flavor computation.

Your maxjetflavor is set to a 4 flavor computation but your proton definition is set to a 5 flavor computation.
This is inconsistent and the matching is bound to fail with those parameters.

You also use a model where the b is massive. (you should use a sm-no_b_mass for a 5 flavor computation)

Cheers,

Olivier

On Jul 8, 2014, at 1:11 PM, Yong Yang <email address hidden> wrote:

> New question #251356 on MadGraph5_aMC@NLO:
> https://answers.launchpad.net/mg5amcnlo/+question/251356
>
> Dear experts,
>
> I have a problem of matching events with pythia, generated from (MG5v1.5.11)
>
> define p u c s d b u~ c~ s~ d~ b~ g
> define j = p
> generate p p > b b~ @0
> add process p p > b b~ j @1
>
> the proc card can be found here:
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor4/proc_card_mg5.dat.txt
>
> for the matching I set the following relevant parameters:
>
> 1 = ickkw ! 0 no matching, 1 MLM, 2 CKKW matching
>
> 20 = ptj ! minimum pt for the jets
> 20 = ptb ! minimum pt for the b
> 4 = maxjetflavor ! Apply b cuts
> #*********************************************************************
> # Jet measure cuts *
> #*********************************************************************
> 20 = xqcut ! minimum kt jet measure between partons
>
>
> the full run_card can be found here:
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor4/run_card.dat.txt
>
> The events generated are then passed to pythia6 for matching with the following relevant parameters:
>
> MAXJETS --> 1
> MINJETS --> 0
> QCUT --> 40
> nQmatch --> 4
>
> The DJR plots can be found here:
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor4/events.treexqcut20qcut40_DJR.gif
>
> which is not smooth at all in the transition region, especially the DJR(1->0)
>
> I have tried to change the QCUT value to 20, and 60 and found no improvement.
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor4/events.treexqcut20qcut20_DJR.gif
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor4/events.treexqcut20qcut60_DJR.gif
>
> I have also tried to change xqcut in run_card.data to 10 and 40 GeV, and try a couple of values of QCUT, and could not find improvement of the DJR plot. for example, for xqut=10
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut10maxjetflavor4/events.treexqcut10qcut10_DJR.gif
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut10maxjetflavor4/events.treexqcut10qcut20_DJR.gif
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut10maxjetflavor4/events.treexqcut10qcut40_DJR.gif
>
> for xqut=40
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut40maxjetflavor4/events.treexqcut40qcut40_DJR.gif
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut40maxjetflavor4/events.treexqcut40qcut60_DJR.gif
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut40maxjetflavor4/events.treexqcut40qcut100_DJR.gif
>
> Do you have any suggestions or comments what can be possibly wrong in the data cards I used, or what can be improved for
> the jet-parton matching for this case?
>
> One more information, before I tried to 4 flavour matching,
> 4 = maxjetflavor ! Apply b cuts
>
> I have tried to set it to be 5 (then nQmatch is set to be 5 as well)
> , but the matching efficiencies are always 0. So I have changed it to 4.
>
>
> Thank you,
> Yong
>
>
>
>
>
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Dear Olivier,

I have tried to use sm-no_b_mass for 5 flavor,
all the data cards can be found here

http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor5/

and you can find the 10k events generated there as well.

for the matching with pythia6,

I have set the b mass to be zero to be consistent with the values used in MG.
     'PMAS(5,1)=0 ! b quark mass',

and the following matching parameters.

MAXJETS --> 1
MINJETS --> 0
QCUT --> 40
nQmatch --> 5

However, I got 0 events matched ( I only run 100 events, because it runs very slow for some reasons
if the b-quark mass is set to be 0, 40 minutes for 100 events)

1********* PYSTAT: Statistics on Number of Events and Cross-sections *********

 ==============================================================================
 I I I I
 I Subprocess I Number of points I Sigma I
 I I I I
 I----------------------------------I----------------------------I (mb) I
 I I I I
 I N:o Type I Generated Tried I I
 I I I I
 ==============================================================================
 I I I I
 I 0 All included subprocesses I 0 100 I 0.000D+00 I
 I 4 User process 1 I 0 21 I 0.000D+00 I
 I 6 User process 0 I 0 79 I 0.000D+00 I
 I I I I
 ==============================================================================

I also tried to set it 4.8 (which is used in CMS)
     'PMAS(5,1)=4.8 ! b quark mass',

it runs much much faster (100 events in few seconds) . but still I get 0 match efficiency

Do you have any ideas?

Thanks,
Yong

Hi,

I have run your files in the latest version of the code (2.1.2)
But not in gridpack mode (such that I can directly run the shower via pythia6)

Here is some comment:
1) The time to create the gridpack is independent of the number of events that you ask inside the run_card
2) Your param_card is not valid for the model that you use. The latest version of the code prevent me to run
with this param_card since a couple of parameter (CKM/mass/yukawa) are different of zero even if the model have them define to zero/diagonal.
So I have change your param_card to be coherent with the model.
3) I have generate 50k, and after matching I have 4108 events (this takes less than a minute to run). Note that the DJR plot didn’t look nice with that choice…
The efficiency for your choice of parameter is very low, so asking for only 100 events doesn’t seem a good idea.

In any case, everything seems fine with the latest version of the code (the only one on which we provide support)

Cheers,

Olivier

On Jul 9, 2014, at 9:56 AM, Yong Yang <email address hidden> wrote:

> Question #251356 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/251356
>
> Status: Answered => Open
>
> Yong Yang is still having a problem:
> Dear Olivier,
>
> I have tried to use sm-no_b_mass for 5 flavor,
> all the data cards can be found here
>
> http://www.physik.uzh.ch/~yangyong/MG/bb1j/xqcut20maxjetflavor5/
>
> and you can find the 10k events generated there as well.
>
> for the matching with pythia6,
>
> I have set the b mass to be zero to be consistent with the values used in MG.
> 'PMAS(5,1)=0 ! b quark mass',
>
> and the following matching parameters.
>
> MAXJETS --> 1
> MINJETS --> 0
> QCUT --> 40
> nQmatch --> 5
>
> However, I got 0 events matched ( I only run 100 events, because it runs very slow for some reasons
> if the b-quark mass is set to be 0, 40 minutes for 100 events)
>
> 1********* PYSTAT: Statistics on Number of Events and Cross-sections
> *********
>
> ==============================================================================
> I I I I
> I Subprocess I Number of points I Sigma I
> I I I I
> I----------------------------------I----------------------------I (mb) I
> I I I I
> I N:o Type I Generated Tried I I
> I I I I
> ==============================================================================
> I I I I
> I 0 All included subprocesses I 0 100 I 0.000D+00 I
> I 4 User process 1 I 0 21 I 0.000D+00 I
> I 6 User process 0 I 0 79 I 0.000D+00 I
> I I I I
> ==============================================================================
>
> I also tried to set it 4.8 (which is used in CMS)
> 'PMAS(5,1)=4.8 ! b quark mass',
>
> it runs much much faster (100 events in few seconds) . but still I get 0
> match efficiency
>
> Do you have any ideas?
>
> Thanks,
> Yong
>
> --
> You received this question notification because you are an answer
> contact for MadGraph5_aMC@NLO.

Hi Olivier,

Thanks for your check. I have tried to use the latest version of the code.

below is what I did:

./bin/mg5_aMC
>mport model sm-no_b_mass
define p u c s d b u~ c~ s~ d~ b~ g
define j = p
generate p p > b b~ @0
add process p p > b b~ j @1
output bb1j

then I go to the directory of bb1j and modified the Card/run_card.dat
 20 = ptj ! minimum pt for the jets
 20 = ptb ! minimum pt for the b
 20 = xqcut ! minimum kt jet measure between partons
 5 = maxjetflavor ! Maximum jet pdg code

(the full card can be found here:
http://www.physik.uzh.ch/~yangyong/MG/bb1j/mg212/run_card.dat.txt)

I also created the Cards/pythia_card.dat, copied from the default one, just add QCUT.

QCUT=40

then I run:
./bin/generate_events -f

the end of the print is:

=== Results Summary for run: run_04 tag: tag_1 ===

     Cross-section : 1.977e+06 +- 1.128e+04 pb
     Nb of events : 1000
     Matched Cross-section : 1.73e+05 +- 1.786e+04 pb
     Nb of events after Matching : 86

So the matching efficiency is close to what you got .

One problem what I found is that when I tried to add the following lines into the Cards/pythia_card.dat
MINJETS=0
MAXJETS=1
SHOWERKT=F

it crashes with pythia.

store_events
INFO: Storing parton level results
INFO: End Parton
reweight -from_cards
decay_events -from_cards
INFO: Running Pythia
Command "generate_events -f" interrupted with error:
ZeroDivisionError : float division by zero
Please report this bug on https://bugs.launchpad.net/madgraph5
More information is found in '/afs/cern.ch/work/y/yangyong/CMSSW/MG/CMSSW_6_2_9/src/MG/madgraph5/bb1j_test/run_05_tag_1_debug.log'.
Please attach this file to your report.
quit
INFO: storring files of Previous run
INFO: Storing Pythia files of Previous run
INFO: Done
INFO:

you can see the log file here:
http://www.physik.uzh.ch/~yangyong/MG/bb1j/mg212/run_05_tag_1_debug.log

Do you have any suggestions ?

Thanks again for your help.

Best,
Yong

Hi yong,

Never realy play with the MINJETS/MAXJETS options (since MG set those automatically) so I'm not an expert on those.

But since you are in 5 flavor, I would have expect that the correct set for your process is
MINJETS=2
MAXJETS=3

Can you try with that option?

Cheers,

Olivier

Hi Olivier,

I have tried, that choice of values will run.

Thanks,
Yong