Drastic parton showering effects at NLO

Asked by Hantian Zhang on 2017-08-13

Hi there,

I am currently running VBF Higgs production in a BSM model at NLO with Pythia8. However, after parton showering, the \Delta \phi of leading and subleading jet distribution has been changed drastically. After showering, it is dominated at \Delta \phi = \pi, which is always back-to-back in x-y plane, however, it is not necessary to be like this, and it is quite different from the \Delta \phi distribution in partion-level (no PS).

Is that possible that some potential problems occur when passing NLO computation to Pythia8 inside MadGraph5 interface? Or somewhere I need to be careful when I use MadGraph and Pythia8 at NLO?

Thank you very much!

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Paolo Torrielli Edit question
Solved by:
Paolo Torrielli
Solved:
2017-08-18
Last query:
2017-08-18
Last reply:
2017-08-17

Dear Hantian,

for Dphi I'd normally expect way more events at Dphi=pi than at Dphi=0,
both for fixed NLO and for NLO+PS, with shower radiation tending to
flatten the distribution (i.e. to fill the Dphi=0 region) with respect to fixed
order. Without looking at plots it is a bit difficult to have an idea of the
drastic difference you see, though.

If you are asking about potential issues linked to the use of the default
Hjj analysis provided with the package, we have used it many times (in
the SM VBF) without noticing suspect features. If you are asking something
else, could you kindly specify what you mean by 'potential problems'?

In any case, a LO+PS run with the same setup, and/or a (N)LO+PS run with
a different shower (e.g. HW++) would be beneficial in order to understand
the origin of these features.

Cheers.
Paolo

Hantian Zhang (hantian01) said : #2

Hi Paolo,

Thank you for your message. However, I checked the VBF Higgs production in SM model by default, the showering effects of Pythia8 does not flatten the distribution. More precisely, the no PS distribution is rather flat, with a bit more events on \Delta \phi = \pi than \Delta \phi =0 (this confirms the results of Maltoni's paper ArXiv: 1311.1829), but after showering the distribution is drastically changed to have pick at \Delta \phi = \pi with very very few events on \Delta \phi = 0.

Then I would guess it is maybe the problem associated to parton shower program...

Best,
Hantian

Hi Hantian,

OK: from the paper you cited, fig. 2, the effect of parton showering over
fixed NLO is rather minimal for Dphi (even smaller than I remembered,
thanks for pointing out that reference).

Are you sure you’re not clustering the decay products of the resonance
into jets at NLO+PS? If this were the case, in the fixed NLO and in the
NLO+PS simulations you would be considering very different jets, that
could explain a dramatic effect.
Moreover, what happens at LO+PS? Do you also see a sizeable distortion?

Thanks.
Cheers.
Paolo

> On 14 Aug 2017, at 17:57, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo,
>
> Thank you for your message. However, I checked the VBF Higgs production
> in SM model by default, the showering effects of Pythia8 does not
> flatten the distribution. More precisely, the no PS distribution is
> rather flat, with a bit more events on \Delta \phi = \pi than \Delta
> \phi =0 (this confirms the results of Maltoni's paper ArXiv: 1311.1829),
> but after showering the distribution is drastically changed to have pick
> at \Delta \phi = \pi with very very few events on \Delta \phi = 0.
>
> Then I would guess it is maybe the problem associated to parton shower
> program...
>
> Best,
> Hantian
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #4

Hi Paolo,

I checked both NLO+PS and LO+PS, PS all have the same drastic effects on Dphi distribution.

For the clustering, I use FastJet in MadAnalysis5 with my own Dphi observable (since MA5 Dphi is not correct in current version).

Would you like to provide more information regards the "clustering the decay products of the resonance into jets at NLO+PS"? Then I could look into details on my simulation to have a check.

Thank you very much!

Best,
Hantian

Hi Hantian,

> Would you like to provide more information regards the "clustering the
> decay products of the resonance into jets at NLO+PS"? Then I could look
> into details on my simulation to have a check.

What I mean is: if, in the showered simulation, the decay chain of
your (Higgs-like) resonance ends up with jets, then the shape of all
jet observables will be very much distorted with respect to fixed order,
where (I guess) you keep the resonance stable.
But this distortion would come only from having (artificially) no jets
from the decay in the fixed-order calculation.

Are you keeping the Higgs stable in your PS simulation (or are you
decaying it to objects that are not clustered into jets)?

Thanks.
Cheers.
Paolo

Hantian Zhang (hantian01) said : #6

Hi Paolo,

Thank you for pointing this. I didn't keep Higgs stable in PS before, which is set to h_stable = F by default. However, after I setting it to be h_stable = T, the drastic PS effects are still there.

Best,
Hantian

Hi Hantian,

OK, strange, I would have thought of an effect of this kind.
Just to be sure, please double check that the events after
showering indeed have stable Higgses (just to make sure that
command in the shower card is correctly understood. Alternatively
you could decay H -> tau tau in the card.).

I’ll think about other sources of this effect. Could you maybe
paste your shower card and run card?

Thanks.
Cheers.
Paolo

> On 14 Aug 2017, at 20:03, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo,
>
> Thank you for pointing this. I didn't keep Higgs stable in PS before,
> which is set to h_stable = F by default. However, after I setting it to
> be h_stable = T, the drastic PS effects are still there.
>
> Best,
> Hantian
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #8

Hi Paolo,

Thank you very much! Here is the shower card and run card:

#***********************************************************************
# MadGraph5_aMC@NLO *
# *
# shower_card.dat aMC@NLO *
# *
# This file is used to set the parameters for the shower. *
# *
# Some notation/conventions: *
# *
# Lines starting with a hash (#) are info or comments *
# *
# mind the format: variable = value # comment *
#***********************************************************************
#
#***********************************************************************
# Shower settings *
#***********************************************************************
# Number of events, jobs, errors, and random seeds *
#***********************************************************************
nevents = -1 # N evts to shower (< 0 = all)
nsplit_jobs = 1 # N jobs to run in parallel (< 100!!)
combine_td = T # combine the topdrawer/HwU files if nsplit_jobs>1
maxprint = 2 # N evts to print in the log
maxerrs = 0.1 # max fraction of errors
rnd_seed = 0 # 1st random seed (0 = default)
rnd_seed2 = 0 # 2nd random seed (0 = default) !ONLY FOR HWERIG6!
#***********************************************************************
# PDFs and non-perturbative modelling *
#***********************************************************************
pdfcode = 0 # 0 = internal, 1 = same as NLO, other = lhaglue
ue_enabled = F # underlying event
hadronize = T # hadronisation on/off !IGNORED BY HERWIG6!
lambda_5 = -1 # Lambda_5 (< 0 = default) !IGNORED BY PYTHIA8!
#***********************************************************************
# Stable or unstable particles *
#***********************************************************************
b_stable = F # set B hadrons stable
pi_stable = T # set pi0's stable
wp_stable = F # set w+'s stable
wm_stable = F # set w-'s stable
z_stable = F # set z0's stable
h_stable = T # set Higgs' stable
tap_stable = F # set tau+'s stable
tam_stable = F # set tau-'s stable
mup_stable = F # set mu+'s stable
mum_stable = F # set mu-'s stable
#***********************************************************************
# Mass of the b quark *
#***********************************************************************
b_mass = -1 # if < 0 = read from SubProcesses/MCmasses_*.inc
#***********************************************************************
# Special settings *
#***********************************************************************
is_4lep = F # T if 4-lepton production !ONLY FOR PYTHIA6!
is_bbar = F # T if bb~ production !ONLY FOR HERWIG6!
#***********************************************************************
# FxFx merging parameters !ONLY FOR PYTHIA8!
#***********************************************************************
Qcut = -1.0 # Merging scale
njmax = 0 # Maximal multiplicity in the merging
#***********************************************************************
# Decay channels *
#***********************************************************************
# Syntax for HERWIG6 *
# DM_I = M > D1 D2 @ BR @ ME *
# corresponding to call to HWMODK(M,BR,ME,D1,D2) *
# I < 100, M is the decaying resonance, D1, D2, ... are the decay *
# products (up to five), BR is the branching ratio and ME is the type *
# of matrix element to be used in the decay. *
# BR's are correctly understood only if they add up to 1, and only if *
# no more than three modes are required for a given resonance. *
# WARNING: the order of decay products in > 2-body decays IS RELEVANT. *
# *
# Syntax for PYTHIA6 *
# DM_I = M > D1 D2 @ BR @ ME *
# WARNING: turning hadronisation off disables top decays *
# WARNING: 1 -> n decays (with n > 2) are handled through a sequence *
# of 1 -> 2 decays. *
# WARNING: entries BR and ME are ignored *
# *
# Syntax for HERWIG++ *
# DM_I = M > D1 D2 @ BR @ ME *
# WARNING: entries BR and ME are ignored *
# *
# Syntax for PYTHIA8 *
# DM_I = M:onIfAny = D1 D2 *
# or similar, according to the offical PYTHIA8 decay syntax, see *
# the online PYTHIA8 manual *
# WARNING: 1 -> n decays (with n > 2) are handled through a sequence *
# of 1 -> 2 decays. *
# *
# Examples *
# Z -> e+ e- or mu+ mu- with BR = 0.5 each, HERWIG6 *
# DM_1 = 23 > -11 11 @ 0.5d0 @ 100
# DM_2 = 23 > -13 13 @ 0.5d0 @ 100
# H -> ta+ ta- with BR = 1, HERWIG6 or HERWIG++ *
# DM_3 = 25 > -15 15 @ 1.0d0 @ 0
# t -> ve e+ b with BR = 1, HERWIG6 or HERWIG++ *
# DM_4 = 6 > 12 -11 5 @ 1d0 @ 100
# t -> ve e+ b with BR = 1, PYTHIA6 *
# DM_5 = 6 > 24 5 @ 1d0 @ 100
# DM_6 = 24 > 12 -11 @ 1d0 @ 100
# W+ -> ve e+, W- -> vm~ mu-, PYTHIA8 *
# DM_1 = 24:onMode = off
# DM_2 = 24:onPosIfAny = 11 12
# DM_3 = 24:onNegIfAny = 13 14
# W+ -> ve e+, W- -> ve~ e- and vm~ mu-, PYTHIA8 *
# DM_1 = 24:onMode = off
# DM_2 = 24:onIfAny = 11 12
# DM_3 = 24:onNegIfAny = 13 14
#***********************************************************************
# Extra libraries/analyses *
#***********************************************************************
# The following lines need to be changed if the user does not want to *
# create a StdHEP/HepMC file, but to directly run an own analysis (to *
# be placed in HWAnalyzer or analogous MCatNLO subfolders). *
# Please use files in those folders as examples. *
# Remember that if your analysis uses hbook or is in the HwU format, *
# you must also add to hbook.o or HwU.o to the ANALYSE list as well. *
#***********************************************************************
extralibs = pythia8 # Extra-libraries (not LHAPDF)
                                 # Default: "stdhep Fmcfio"
                                 # PYTHIA > 8.200 may require library dl
extrapaths = ../lib /app/cern/pythia_8226/lib # Path to the extra-libraries
                                 # Default: "../lib"
INCLUDEPATHS = # Path to header files needed by c++
                                 # Dir names separated by white spaces
ANALYSE = # User's analysis and histogramming
                                 # routines (please use .o as extension
                                 # and use spaces to separate files)

#***********************************************************************
# MadGraph5_aMC@NLO *
# *
# run_card.dat aMC@NLO *
# *
# This file is used to set the parameters of the run. *
# *
# Some notation/conventions: *
# *
# Lines starting with a hash (#) are info or comments *
# *
# mind the format: value = variable ! comment *
# *
# Some of the values of variables can be list. These can either be *
# comma or space separated. *
#***********************************************************************
#
#*******************
# Running parameters
#*******************
#
#***********************************************************************
# Tag name for the run (one word) *
#***********************************************************************
  tag_1 = run_tag ! name of the run
#***********************************************************************
# Number of LHE events (and their normalization) and the required *
# (relative) accuracy on the Xsec. *
# These values are ignored for fixed order runs *
#***********************************************************************
 50000 = nevents ! Number of unweighted events requested
 -1.0 = req_acc ! Required accuracy (-1=auto determined from nevents)
 -1 = nevt_job! Max number of events per job in event generation.
                 ! (-1= no split).
#***********************************************************************
# Normalize the weights of LHE events such that they sum or average to *
# the total cross section *
#***********************************************************************
 average = event_norm ! average or sum
#***********************************************************************
# Number of points per itegration channel (ignored for aMC@NLO runs) *
#***********************************************************************
 0.01 = req_acc_FO ! Required accuracy (-1=ignored, and use the
                     ! number of points and iter. below)
# These numbers are ignored except if req_acc_FO is equal to -1
 5000 = npoints_FO_grid ! number of points to setup grids
 4 = niters_FO_grid ! number of iter. to setup grids
 10000 = npoints_FO ! number of points to compute Xsec
 6 = niters_FO ! number of iter. to compute Xsec
#***********************************************************************
# Random number seed *
#***********************************************************************
 0 = iseed ! rnd seed (0=assigned automatically=default))
#***********************************************************************
# Collider type and energy *
#***********************************************************************
 1 = lpp1 ! beam 1 type (0 = no PDF)
 1 = lpp2 ! beam 2 type (0 = no PDF)
 6500.0 = ebeam1 ! beam 1 energy in GeV
 6500.0 = ebeam2 ! beam 2 energy in GeV
#***********************************************************************
# PDF choice: this automatically fixes also alpha_s(MZ) and its evol. *
#***********************************************************************
 nn23nlo = pdlabel ! PDF set
 244600 = lhaid ! If pdlabel=lhapdf, this is the lhapdf number. Only
              ! numbers for central PDF sets are allowed. Can be a list;
              ! PDF sets beyond the first are included via reweighting.
#***********************************************************************
# Include the NLO Monte Carlo subtr. terms for the following parton *
# shower (HERWIG6 | HERWIGPP | PYTHIA6Q | PYTHIA6PT | PYTHIA8) *
# WARNING: PYTHIA6PT works only for processes without FSR!!!! *
#***********************************************************************
  PYTHIA8 = parton_shower
  1.0 = shower_scale_factor ! multiply default shower starting
                                  ! scale by this factor
#***********************************************************************
# Renormalization and factorization scales *
# (Default functional form for the non-fixed scales is the sum of *
# the transverse masses divided by two of all final state particles *
# and partons. This can be changed in SubProcesses/set_scales.f or via *
# dynamical_scale_choice option) *
#***********************************************************************
 False = fixed_ren_scale ! if .true. use fixed ren scale
 False = fixed_fac_scale ! if .true. use fixed fac scale
 91.118 = muR_ref_fixed ! fixed ren reference scale
 91.118 = muF_ref_fixed ! fixed fact reference scale
 -1 = dynamical_scale_choice ! Choose one (or more) of the predefined
           ! dynamical choices. Can be a list; scale choices beyond the
           ! first are included via reweighting
 1.0 = muR_over_ref ! ratio of current muR over reference muR
 1.0 = muF_over_ref ! ratio of current muF over reference muF
#***********************************************************************
# Reweight variables for scale dependence and PDF uncertainty *
#***********************************************************************
 1.0, 2.0, 0.5 = rw_rscale ! muR factors to be included by reweighting
 1.0, 2.0, 0.5 = rw_fscale ! muF factors to be included by reweighting
 True = reweight_scale ! Reweight to get scale variation using the
            ! rw_rscale and rw_fscale factors. Should be a list of
            ! booleans of equal length to dynamical_scale_choice to
            ! specify for which choice to include scale dependence.
 False = reweight_PDF ! Reweight to get PDF uncertainty. Should be a
            ! list booleans of equal length to lhaid to specify for
            ! which PDF set to include the uncertainties.
#***********************************************************************
# Store reweight information in the LHE file for off-line model- *
# parameter reweighting at NLO+PS accuracy *
#***********************************************************************
 False = store_rwgt_info ! Store info for reweighting in LHE file
#***********************************************************************
# ickkw parameter: *
# 0: No merging *
# 3: FxFx Merging - WARNING! Applies merging only at the hard-event *
# level. After showering an MLM-type merging should be applied as *
# well. See http://amcatnlo.cern.ch/FxFx_merging.htm for details. *
# 4: UNLOPS merging (with pythia8 only). No interface from within *
# MG5_aMC available, but available in Pythia8. *
# -1: NNLL+NLO jet-veto computation. See arxiv:1412.8408 [hep-ph]. *
#***********************************************************************
 0 = ickkw
#***********************************************************************
#
#***********************************************************************
# BW cutoff (M+/-bwcutoff*Gamma). Determines which resonances are *
# written in the LHE event file *
#***********************************************************************
 15.0 = bwcutoff
#***********************************************************************
# Cuts on the jets. Jet clustering is performed by FastJet. *
# - When matching to a parton shower, these generation cuts should be *
# considerably softer than the analysis cuts. *
# - More specific cuts can be specified in SubProcesses/cuts.f *
#***********************************************************************
  -1.0 = jetalgo ! FastJet jet algorithm (1=kT, 0=C/A, -1=anti-kT)
  0.4 = jetradius ! The radius parameter for the jet algorithm
 30.0 = ptj ! Min jet transverse momentum
 4.5 = etaj ! Max jet abs(pseudo-rap) (a value .lt.0 means no cut)
#***********************************************************************
# Cuts on the charged leptons (e+, e-, mu+, mu-, tau+ and tau-) *
# More specific cuts can be specified in SubProcesses/cuts.f *
#***********************************************************************
  0.0 = ptl ! Min lepton transverse momentum
 -1.0 = etal ! Max lepton abs(pseudo-rap) (a value .lt.0 means no cut)
  0.0 = drll ! Min distance between opposite sign lepton pairs
  0.0 = drll_sf ! Min distance between opp. sign same-flavor lepton pairs
  0.0 = mll ! Min inv. mass of all opposite sign lepton pairs
  0.0 = mll_sf ! Min inv. mass of all opp. sign same-flavor lepton pairs
#***********************************************************************
# Photon-isolation cuts, according to hep-ph/9801442. When ptgmin=0, *
# all the other parameters are ignored. *
# More specific cuts can be specified in SubProcesses/cuts.f *
#***********************************************************************
  0.0 = ptgmin ! Min photon transverse momentum
 -1.0 = etagamma ! Max photon abs(pseudo-rap)
  0.4 = R0gamma ! Radius of isolation code
  1.0 = xn ! n parameter of eq.(3.4) in hep-ph/9801442
  1.0 = epsgamma ! epsilon_gamma parameter of eq.(3.4) in hep-ph/9801442
 True = isoEM ! isolate photons from EM energy (photons and leptons)
#***********************************************************************
# For aMCfast+APPLGRID use in PDF fitting (http://amcfast.hepforge.org)*
#***********************************************************************
 0 = iappl ! aMCfast switch (0=OFF, 1=prepare grids, 2=fill grids)
#***********************************************************************

marco zaro (marco-zaro) said : #9

Hi Hantian,
not much to add here.
Are you sure that you are applying the same cuts at the analysis level at NLOPS and at fNLO?
also, what are the event-generation cuts that you use to generate the sample to be showered?
Cheers,
Marco

Hantian Zhang (hantian01) said : #10

Hi Marco,

Thank you for your message, I am not sure if I understand your question properly, so here is how I did my simulation and analysis:

I employ the VBF cuts in the run card with, etaj < 4.5 and ptj>30.0 GeV, also anti-kt FastJet algorithm with radius 0.4.
Then I proceed the NLO+PS with Pythia8 inside MadGraph.
The resulting .hepmc file was imported into MadAnalysis5 in reconstructed mode with additional cuts of invariant jets mass mj1j2>500 GeV. The MadAnalysis invoke FastJet for jet clustering automatically, and then I got the distributions.

Best,
Hantian

Hi Hantian,

for a (N)LO+PS run, the cuts you use while generating events
must be significantly looser than the ones you use in the
analysis, in order not to bias the result by excluding (in
generation) events that would be moved by the shower in the
accepted region.

If you use the run card below for NLO+PS, did you make sure
you’re not producing such a bias? Could you try and generate
without cut in eta, and with a smaller cut in ptj, in order
to check if the showered results are the same?

Thanks.
Cheers.
Paolo

> On 14 Aug 2017, at 22:22, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Marco,
>
> Thank you for your message, I am not sure if I understand your question
> properly, so here is how I did my simulation and analysis:
>
> I employ the VBF cuts in the run card with, etaj < 4.5 and ptj>30.0 GeV, also anti-kt FastJet algorithm with radius 0.4.
> Then I proceed the NLO+PS with Pythia8 inside MadGraph.
> The resulting .hepmc file was imported into MadAnalysis5 in reconstructed mode with additional cuts of invariant jets mass mj1j2>500 GeV. The MadAnalysis invoke FastJet for jet clustering automatically, and then I got the distributions.
>
> Best,
> Hantian
>
> --
> You received this question notification because you are subscribed to
> the question.

Hantian Zhang (hantian01) said : #12

Hi Paolo,

I checked that with loose cut, and it does not improve the results. This is what I would expect, since pt and eta cut should not sizeably change the Dphi distribution. Thanks.

Best,
Hantian

Hantian Zhang (hantian01) said : #13

Hi Paolo,

I would like to change to Herwig7 for some testing. I assign the Herwig7 path to the mg5_configuration.txt with , hwpp_path, thepeg_path, hepmc_path, to be the Herwig7 path, and choose HERWIGPP in the run card.
However, the showering process output with error:
IOError : [Errno 2] No such file or directory: '/MG5_aMC_v2_5_5_ForTest/bin/VBF_SMmodel/MCatNLO/RUN_HERWIGPP_3/events.hepmc.gz'

Would you like to give some advice for this issue? Thank you very much.

Best,
Hantian

Hi Hantian,

if you change mg5_configuration.txt after having already
generated the process, the changes won't apply.

Please try and change also the file
<Process>/Cards/amcatnlo_configuration.txt

Cheers.
Paolo

> On 15 Aug 2017, at 16:37, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo,
>
> I would like to change to Herwig7 for some testing. I assign the Herwig7 path to the mg5_configuration.txt with , hwpp_path, thepeg_path, hepmc_path, to be the Herwig7 path, and choose HERWIGPP in the run card.
> However, the showering process output with error:
> IOError : [Errno 2] No such file or directory: '/MG5_aMC_v2_5_5_ForTest/bin/VBF_SMmodel/MCatNLO/RUN_HERWIGPP_3/events.hepmc.gz'
>
> Would you like to give some advice for this issue? Thank you very much.
>
> Best,
> Hantian
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #15

Hi Paolo,

It does not resolve the Herwig problem, here I attach the debug output, would you like to take a look? Thank you

#************************************************************
#* MadGraph5_aMC@NLO *
#* *
#* * * *
#* * * * * *
#* * * * * 5 * * * * *
#* * * * * *
#* * * *
#* *
#* *
#* VERSION 2.5.5 2017-05-26 *
#* *
#* The MadGraph5_aMC@NLO Development Team - Find us at *
#* https://server06.fynu.ucl.ac.be/projects/madgraph *
#* and *
#* http://amcatnlo.cern.ch *
#* *
#************************************************************
#* *
#* Command File for aMCatNLO *
#* *
#* run as ./bin/aMCatNLO.py filename *
#* *
#************************************************************
launch auto --name SM_Herwig
Traceback (most recent call last):
  File "/disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/madgraph/interface/extended_cmd.py", line 1430, in onecmd
    return self.onecmd_orig(line, **opt)
  File "/disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/madgraph/interface/extended_cmd.py", line 1384, in onecmd_orig
    return func(arg, **opt)
  File "/disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/madgraph/interface/amcatnlo_run_interface.py", line 1246, in do_launch
    self.run_mcatnlo(evt_file, options)
  File "/disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/madgraph/interface/amcatnlo_run_interface.py", line 3457, in run_mcatnlo
    files.mv(os.path.join(rundir, 'events.%s.gz' % ext), hep_file)
  File "/disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/madgraph/iolibs/files.py", line 172, in mv
    shutil.move(path1, path2)
  File "/usr/lib64/python2.7/shutil.py", line 302, in move
    copy2(src, real_dst)
  File "/usr/lib64/python2.7/shutil.py", line 130, in copy2
    copyfile(src, dst)
  File "/usr/lib64/python2.7/shutil.py", line 82, in copyfile
    with open(src, 'rb') as fsrc:
IOError: [Errno 2] No such file or directory: '/disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/bin/VBF_SM/MCatNLO/RUN_HERWIGPP_3/events.hepmc.gz'
Related File: /disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/bin/VBF_SM/MCatNLO/RUN_HERWIGPP_3/events.hepmc.gz
Value of current Options:
              text_editor : None
      notification_center : True
                    pjfry : None
       cluster_local_path : None
       group_subprocesses : Auto
ignore_six_quark_processes : False
    loop_optimized_output : True
    cluster_status_update : (600, 30)
         fortran_compiler : None
               hepmc_path : /disk/scratch/herwig/herwig_7.1.1
                  collier : /disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/HEPTools/lib
              auto_update : 7
             pythia8_path : /app/cern/pythia_8226
                hwpp_path : /disk/scratch/herwig/herwig_7.1.1
low_mem_multicore_nlo_generation : False
                    golem : None
          pythia-pgs_path : None
                  td_path : None
             delphes_path : /disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/Delphes
              thepeg_path : /disk/scratch/herwig/herwig_7.1.1
             cluster_type : condor
        madanalysis5_path : None
      exrootanalysis_path : None
                      OLP : MadLoop
                 applgrid : applgrid-config
               eps_viewer : None
                  fastjet : fastjet-config
                 run_mode : 2
              web_browser : None
   automatic_html_opening : False
        cluster_temp_path : None
             cluster_size : 100
            cluster_queue : None
             syscalc_path : None
         madanalysis_path : None
                   lhapdf : /disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/HEPTools/lhapdf6/bin/lhapdf-config
             stdout_level : 20
                  nb_core : 48
            f2py_compiler : None
                    ninja : /app/theorie/ninja_1.1.0/lib64
                  amcfast : amcfast-config
       cluster_retry_wait : 300
      output_dependencies : external
           crash_on_error : False
mg5amc_py8_interface_path : /disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/HEPTools/MG5aMC_PY8_interface
         loop_color_flows : False
                  samurai : None
         cluster_nb_retry : 1
                 mg5_path : /disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest
                  timeout : 60
                    gauge : unitary
      complex_mass_scheme : False
             cpp_compiler : None
   max_npoint_for_channel : 0

marco zaro (marco-zaro) said : #16

Hi Hantian, what is the content of this dir?
disk/data7/ttp/zhangha/MG5_aMC_v2_5_5_ForTest/bin/VBF_SM/MCatNLO/RUN_HERWIGPP_3/
cheers,

Marco

Hantian Zhang (hantian01) said : #17

Hi Marco,

It contains:
events.lhe Herwig MCATNLO_HERWIGPP_EXE MCATNLO_HERWIGPP_input mcatnlo_run.log shower_card.dat shower.sh

Best,
Hantian

Hi Hantian, Marco,

I think we have to solve one issue at a time, in order to converge.
Let’s please first come back to the (SM) VBF (N)LO+PY8 problem.

If I understand correctly, you use the provided analysis in the
FixedOrderAnalysis folder to plot fixed NLO quantities, while you
use MadAnalysis5 to analyse the samples after shower.

As you are using two different tools, the risk of inconsistencies
or simply different conventions in the naive definition of observables
is non negligible.

I assume that at fixed NLO you reproduced quantitatively the results
in 1311.1829, so the problem comes from the showered case.

Could you check with the VBF analyses provided in the PY8Analyzer folder?
These can be called directly from the shower_card, and produce output
files (.HwU or .top) in the Events/run_* folder. These output files
can be compared to the analogous .HwU or .top ones you produced at fNLO.

You wisely chose HW7 to cross check the PY8 behaviour. However,
since this is not technically working for the moment in your setup
(and we’ll soon after try and address this problem), could you instead
cross-check the PY8 behaviour with HW6 and/or PY6 (Q-ordered), which
are effortlessly run?

Once these checks are done, and we hopefully have understood the physical
nature of the VBF problem, I propose to address the HW7 issue.

Cheers.
Paolo

> On 16 Aug 2017, at 16:33, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Marco,
>
> It contains:
> events.lhe Herwig MCATNLO_HERWIGPP_EXE MCATNLO_HERWIGPP_input mcatnlo_run.log shower_card.dat shower.sh
>
> Best,
> Hantian
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #19

Hi Paolo,

Thank you for the suggestions. For analysis, I use MadAnalysis5 for all analysis ( NLO/LO and NLO/LO + PS ), and the observable is correct, so this possibility could be excluded I think. I will try HW6 or PY6 to have tests.

Best,
Hantian

Hi Hantian,

thanks for clarifying, but how do you use MadAnalysis5 at fixed NLO
(i.e. the commands you type, and cards you edit)?
Could you confirm you reproduce the fixed order results of that
1311.1829 paper?

Thanks.
Cheers.
Paolo

> On 16 Aug 2017, at 17:18, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo,
>
> Thank you for the suggestions. For analysis, I use MadAnalysis5 for all
> analysis ( NLO/LO and NLO/LO + PS ), and the observable is correct, so
> this possibility could be excluded I think. I will try HW6 or PY6 to
> have tests.
>
> Best,
> Hantian
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #21

Hi Paolo,

Yes, I reproduce the NLO (no PS) SM VBF results of that paper 1311.1829, but also similar transversal and CP-odd structure distributions in our BSM model, so this procedure should be correct.

For the usage of MadAnalysis, I entered the expert mode and modified the Dphi observable, which is validated in NLO (no PS) case.

Best,
Hantian

Hantian Zhang (hantian01) said : #22

Hi Paolo,

May I ask if HW6 and PY6 is installed together with MG5? As I did not find the place to put path in configure files, but I encountered compilation issues with HW6 and PY6, and there is almost no material to be found on website regarding HW6/PY6 with MG5.

Thanks!

Best,
Hantian

Hantian Zhang (hantian01) said : #23

Hi Paolo,

I have test it with HW6 and PY6Q, however, the drastic PS effect is still there....

Best,
Hantian

marco zaro (marco-zaro) said : #24

Dear Hantian,
can you please check if the effect is there with the analysis that is shipped within MG5_aMC?
you should add it to the shower_card, like this (for HW6)

EXTRALIBS = fastjet # Extra-libraries (not LHAPDF)
                                 # Default: "stdhep Fmcfio"
                                 # PYTHIA > 8.200 may require library dl
EXTRAPATHS = ../lib # Path to the extra-libraries
                                 # Default: "../lib"
INCLUDEPATHS = # Path to header files needed by c++
                                 # Dir names separated by white spaces
ANALYSE = mcatnlo_hwan_pp_hjj.o mcatnlo_hbook_gfortran8.o myfastjetfortran.o

then you have the .top file which contains the histograms in a human-readable format, you should be able to check if the effect are still there.
Note that the VBF cuts are defined inside MCatNLO/srcHerwig/mcatnlo_hwan_pp_hjj.f

Finally, how do you do the plot at fixed order (NLO without PS)? do you take the lhe file at parton level or you do a fixed-order run?

cheers,

Marco

On 16 Aug 2017, at 22:43, Hantian Zhang <email address hidden> wrote:

> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo,
>
> I have test it with HW6 and PY6Q, however, the drastic PS effect is
> still there....
>
> Best,
> Hantian
>
> --
> You received this question notification because you are subscribed to
> the question.

Hantian Zhang (hantian01) said : #25

Hi Marco,

Thank you! I have checked the output with HW6 and your commands, the effects are still there in the Dphi 1-j2 output in the .top file. Except I did not find this file: MCatNLO/srcHerwig/mcatnlo_hwan_pp_hjj.f

For my plot, I take lhe file at parton-level, not from a fixed-order run.

Best,
Hantian

Hi Hantian,

sorry, there is a fundamental thing I don’t understand.
Are you plotting the .lhe files before showering?

These results, at NLO, are *non-physical*, so they
cannot be plotted to understand the physical behaviour
of the observables.

The analysis Marco is mentioning is in HWAnalyzer.
This will produce a .top file. Please, pay particular
attention at he definition of VBF cuts, which you may
want edit.

Analogously, by setting

ANALYSE = hw6an_HwU_pp_hjj.o HwU.o myfastjetfortran.o

you will produce a .HwU file that can be processed using
gnuplot.

Cheers.
Paolo

> On 16 Aug 2017, at 23:33, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Marco,
>
> Thank you! I have checked the output with HW6 and your commands, the
> effects are still there in the Dphi 1-j2 output in the .top file. Except
> I did not find this file: MCatNLO/srcHerwig/mcatnlo_hwan_pp_hjj.f
>
> For my plot, I take lhe file at parton-level, not from a fixed-order
> run.
>
> Best,
> Hantian
>
> --
> You received this question notification because you are subscribed to
> the question.

Hantian Zhang (hantian01) said : #27

Hi Paolo and Marco,

Thank you for the patience and explanation. I tried fixed order run (with analysis_td_pp_hjj.f), and reproduced the SM VBF Dphi distribution as that in 1311.1829.

In the NLO+PS PY6Q run (with mactnlo_pyan_pp_hjj.f), I also reproduced the SM VBF Dphi distribution.

However, in NLO+PS HW6 run (with mactnlo_hwan_pp_hjj.f and hw6an_HwU_pp_hjj.f), the drastic PS effects were still there.

Best,
Hantian

marco zaro (marco-zaro) said : #28

Hi Hantian,
good to know things are better now
for HW6 (Paolo please confirm) setting the higgs stable in the card does not work, so you have mostly higgs->bb, and thus extra jets
Rather, you have to change the source code, in particular
srcHerwig/madfks_hwdriver.f, line 50
from
      IPROC=-18000

to
      IPROC=-18017
where the ’17’ forces the decay to taus.
Finally, in the shower_card, set the tau+ and tau- stable

let us know…
cheers,

Marco
On 17 Aug 2017, at 16:27, Hantian Zhang <email address hidden> wrote:

> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo and Marco,
>
> Thank you for the patience and explanation. I tried fixed order run
> (with analysis_td_pp_hjj.f), and reproduced the SM VBF Dphi distribution
> as that in 1311.1829.
>
> In the NLO+PS PY6Q run (with mactnlo_pyan_pp_hjj.f), I also reproduced
> the SM VBF Dphi distribution.
>
> However, in NLO+PS HW6 run (with mactnlo_hwan_pp_hjj.f and
> hw6an_HwU_pp_hjj.f), the drastic PS effects were still there.
>
> Best,
> Hantian
>
> --
> You received this question notification because you are subscribed to
> the question.

Dear Hantian,
thanks for your check.

> In the NLO+PS PY6Q run (with mactnlo_pyan_pp_hjj.f), I also reproduced
> the SM VBF Dphi distribution.
>
> However, in NLO+PS HW6 run (with mactnlo_hwan_pp_hjj.f and
> hw6an_HwU_pp_hjj.f), the drastic PS effects were still there.

This is half good news. The only difference (that comes to my mind)
between HW6 and PY6 in this case is that HW6 would ignore the command
where you set the H stable in the shower card (you just cannot set
H, W, and Z stable in HW6), so I guess the jets you plot in that case
get contributions from the Higgs decays.

Could you please, for HW6, let the Higgs decay to taus, and
set the taus as stable particles (all in the shower card)?

Cheers,
Paolo

Hi Marco, Hantian,
our wires crossed..

> srcHerwig/madfks_hwdriver.f, line 50
> from
> IPROC=-18000
> to
> IPROC=-18017
this should not be necessary, in principle one should
be able to have that behaviour only editing shower_card.

In case this doesn’t work (namely if the Higgses in the
HW6 event record in MCatNLO/RUN*/mcatnlo_run.log decay not
just to taus), then Marco’s suggestion is of course valuable.

Cheers.
Paolo

> where the ’17’ forces the decay to taus.
> Finally, in the shower_card, set the tau+ and tau- stable
>
> let us know…
> cheers,
>
> Marco
> On 17 Aug 2017, at 16:27, Hantian Zhang <email address hidden> wrote:
>
>> Question #655841 on MadGraph5_aMC@NLO changed:
>> https://answers.launchpad.net/mg5amcnlo/+question/655841
>>
>> Hantian Zhang posted a new comment:
>> Hi Paolo and Marco,
>>
>> Thank you for the patience and explanation. I tried fixed order run
>> (with analysis_td_pp_hjj.f), and reproduced the SM VBF Dphi distribution
>> as that in 1311.1829.
>>
>> In the NLO+PS PY6Q run (with mactnlo_pyan_pp_hjj.f), I also reproduced
>> the SM VBF Dphi distribution.
>>
>> However, in NLO+PS HW6 run (with mactnlo_hwan_pp_hjj.f and
>> hw6an_HwU_pp_hjj.f), the drastic PS effects were still there.
>>
>> Best,
>> Hantian
>>
>> --
>> You received this question notification because you are subscribed to
>> the question.
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #31

Hi Marco and Paolo,

The Dphi distribution is now correct with the following setting with HW6:
h_stable = F
tap_stable = T
tam_stable = T
DM_3 = 25 > -15 15 @ 1.0d0 @ 0

Thank you very much!

Best,
Hantian

Hi Hantian,

great. The next step would thus be to try the analogous
analysis with PY8..

Let us know.
Cheers.
Paolo

> On 17 Aug 2017, at 17:37, Hantian Zhang <email address hidden> wrote:
>
> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Marco and Paolo,
>
> The Dphi distribution is now correct with the following setting with HW6:
> h_stable = F
> tap_stable = T
> tam_stable = T
> DM_3 = 25 > -15 15 @ 1.0d0 @ 0
>
> Thank you very much!
>
> Best,
> Hantian
>
> --
> You received this question notification because you are assigned to this
> question.

Hantian Zhang (hantian01) said : #33

Hi Paolo,

PY8 also works by setting Higgs to be stable with py8an_HwU_pp_hjj.f

Best,
Hantian

Hantian Zhang (hantian01) said : #34

Thanks Paolo Torrielli, that solved my question.

Hantian Zhang (hantian01) said : #35

Hi Paolo and Marco,

May I ask one more question regards usage of mcatnlo_pyan_pp_hjj.f for generating 2D histogram data in .top file?
I inserted the following codes in mcatnlo_pyan_pp_hjj.f

in PYABEG subroutine:
         call mbook2(l+ 55,'2D Dphi - j1 pT' //cc(i)//weights_info(kk),pi/8.d0,0d0,pi,100.d0,0d0,800d0)

in PYAEND subroutine:
      OPEN(UNIT=99,FILE='PYTVBF2D.top',STATUS='UNKNOWN')
      XNORM=1.D0/IEVT
      DO I=1,NPL
        CALL MFINAL32(I)
        CALL MCOPY2(I,I+NPL)
        CALL MOPERA2(I+NPL,'F',I+NPL,I+NPL,(XNORM),0.D0)
        CALL MFINAL32(I+NPL)
      ENDDO
      do kk=1,nwgt_analysis
      do i=1,2
        l=(kk-1)*110+(i-1)*55
         call mtop2(NPL+l+ 55,'2D Dphi - j1 pT ',' ','LOG')
      enddo
      enddo
      CLOSE(99)

in PYANAL subroutine
            call mfill2(l+ 55,Dphij1j2,ptj1,www(kk))

But the second .top file generated is empty. Do I made any mistake here or where can I find the output? Thank you!

Best,
Hantian

marco zaro (marco-zaro) said : #36

Dear Hantinan,
I have never used mbook2 (bi-dimensional histograms) plus the whole hbook code is very old, so I cannot really help on this…
One alternative thing is to book 8 1-d histograms, each for one of the phi bins
Cheers,

Marco

On 19 Aug 2017, at 20:23, Hantian Zhang <email address hidden> wrote:

> Question #655841 on MadGraph5_aMC@NLO changed:
> https://answers.launchpad.net/mg5amcnlo/+question/655841
>
> Hantian Zhang posted a new comment:
> Hi Paolo and Marco,
>
> May I ask one more question regards usage of mcatnlo_pyan_pp_hjj.f for generating 2D histogram data in .top file?
> I inserted the following codes in mcatnlo_pyan_pp_hjj.f
>
> in PYABEG subroutine:
> call mbook2(l+ 55,'2D Dphi - j1 pT' //cc(i)//weights_info(kk),pi/8.d0,0d0,pi,100.d0,0d0,800d0)
>
> in PYAEND subroutine:
> OPEN(UNIT=99,FILE='PYTVBF2D.top',STATUS='UNKNOWN')
> XNORM=1.D0/IEVT
> DO I=1,NPL
> CALL MFINAL32(I)
> CALL MCOPY2(I,I+NPL)
> CALL MOPERA2(I+NPL,'F',I+NPL,I+NPL,(XNORM),0.D0)
> CALL MFINAL32(I+NPL)
> ENDDO
> do kk=1,nwgt_analysis
> do i=1,2
> l=(kk-1)*110+(i-1)*55
> call mtop2(NPL+l+ 55,'2D Dphi - j1 pT ',' ','LOG')
> enddo
> enddo
> CLOSE(99)
>
> in PYANAL subroutine
> call mfill2(l+ 55,Dphij1j2,ptj1,www(kk))
>
> But the second .top file generated is empty. Do I made any mistake here
> or where can I find the output? Thank you!
>
> Best,
> Hantian
>
> --
> You received this question notification because you are subscribed to
> the question.

Hantian Zhang (hantian01) said : #37

Hi Marco,

Thank you, then I would just book 8 histograms.

Best,
Hantian