MadGraph5_aMC@NLO

When generating p p > j j [QCD] process at NLO, the total cross-section is negative

Asked by Matthew De Angelis

I am attempting to conduct a fixed-order calculation at NLO of the process p p > j j, with the aim of confirming a gaps between jets analysis using NLOJET++. Please see below, the run_card.dat file:

------------------------------------------------------------------------------------------------------------------------------------------------

#***********************************************************************
# 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 *
#***********************************************************************
#
#*******************
# 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 *
#***********************************************************************
 10000 = 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)
 7000.0 = ebeam1 ! beam 1 energy in GeV
 7000.0 = ebeam2 ! beam 2 energy in GeV
#***********************************************************************
# PDF choice: this automatically fixes also alpha_s(MZ) and its evol. *
#***********************************************************************
 cteq6_l = pdlabel ! PDF set
 244600 = lhaid ! if pdlabel=lhapdf, this is the lhapdf number
#***********************************************************************
# 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!!!! *
#***********************************************************************
  HERWIG6 = 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 of all final state particles and partons. This *
# can be changed in SubProcesses/set_scales.f) *
#***********************************************************************
 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 = muF1_ref_fixed ! fixed fact reference scale for pdf1
 91.118 = muF2_ref_fixed ! fixed fact reference scale for pdf2
 -1 = dynamical_scale_choice ! Choose one of the preselected dynamical choices
#***********************************************************************
# Renormalization and factorization scales (advanced and NLO options) *
#***********************************************************************
 False = fixed_QES_scale ! if .true. use fixed Ellis-Sexton scale
 91.118 = QES_ref_fixed ! fixed Ellis-Sexton reference scale
 1.0 = muR_over_ref ! ratio of current muR over reference muR
 1.0 = muF1_over_ref ! ratio of current muF1 over reference muF1
 1.0 = muF2_over_ref ! ratio of current muF2 over reference muF2
 1.0 = QES_over_ref ! ratio of current QES over reference QES
#***********************************************************************
# Reweight flags to get scale dependence and PDF uncertainty *
# For scale dependence: factor rw_scale_up/down around central scale *
# For PDF uncertainty: use LHAPDF with supported set *
#***********************************************************************
 True = reweight_scale ! reweight to get scale dependence
 0.5 = rw_Rscale_down ! lower bound for ren scale variations
 2.0 = rw_Rscale_up ! upper bound for ren scale variations
 0.5 = rw_Fscale_down ! lower bound for fact scale variations
 2.0 = rw_Fscale_up ! upper bound for fact scale variations
 False = reweight_PDF ! reweight to get PDF uncertainty
 244601 = PDF_set_min ! First of the error PDF sets
 244700 = PDF_set_max ! Last of the error PDF sets
#***********************************************************************
# 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) *
#***********************************************************************
 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)
 0.6 = jetradius ! The radius parameter for the jet algorithm
 20.0 = ptj ! Min jet transverse momentum
 4.4 = 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 gen 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
  30.0 = 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 *
#***********************************************************************
  20.0 = 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)
#***********************************************************************
# Maximal PDG code for quark to be considered a jet when applying cuts.*
# At least all massless quarks of the model should be included here. *
#***********************************************************************
 4 = maxjetflavor
#***********************************************************************
# For aMCfast+APPLGRID use in PDF fitting (http://amcfast.hepforge.org)*
#***********************************************************************
 0 = iappl ! aMCfast switch (0=OFF, 1=prepare APPLgrids, 2=fill grids)
#***********************************************************************

------------------------------------------------------------------------------------------------------------------------------------------------

After running launch DIJETPP -p, I am presented with an events file with a negative total cross-section as seen in the summary/res files pasted below:

------------------------------------------------------------------------------------------------------------------------------------------------

      Summary:
      Process p p > j j [QCD]
      Run at p-p collider (7000.0 + 7000.0 GeV)
      Total cross-section: -1.695e+08 +- 2.1e+07 pb
      Ren. and fac. scale uncertainty: +-181.0% --431.9%
      Number of events generated: 10000
      Parton shower to be used: HERWIG6
      Fraction of negative weights: 0.53
      Total running time : 22m 6s

------------------------------------------------------------------------------------------------------------------------------------------------

Cross-section per integration channel:
            P0_gg_gg 5 -1.74845983e+08 1.8413e+07 -10.5310%
            P0_gu_ug 3 -5.23742925e+06 6.5017e+06 -124.1383%
            P0_ug_ug 2 -7.72242764e+06 4.7092e+06 -60.9808%
          P0_gux_uxg 3 -7.52997347e+05 3.0528e+06 -405.4198%
          P0_uxg_uxg 2 -1.99466775e+06 2.7835e+06 -139.5457%
            P0_gu_ug 2 2.65532390e+06 4.1022e+05 15.4491%
           P0_gg_uux 3 3.70593329e+06 3.9187e+05 10.5741%
            P0_uc_uc 1 4.99998486e+05 3.7583e+05 75.1666%
           P0_gg_uux 2 4.16274620e+06 3.4260e+05 8.2302%
            P0_ug_ug 3 2.07958035e+06 2.9851e+05 14.3545%
          P0_ucx_ucx 1 5.00276732e+05 2.2538e+05 45.0517%
          P0_uxc_uxc 1 5.23612439e+05 2.0284e+05 38.7387%
            P0_uu_uu 1 -3.90250856e+04 1.9053e+05 -488.2360%
          P0_gux_uxg 2 1.15811344e+06 1.6508e+05 14.2542%
        P0_uxcx_uxcx 1 1.57840589e+05 1.3915e+05 88.1561%
          P0_uxg_uxg 3 1.23265667e+06 1.1019e+05 8.9391%
          P0_uux_uux 2 2.06853696e+05 9.0788e+04 43.8901%
            P0_gu_ug 1 4.32575522e+05 8.9652e+04 20.7251%
            P0_ug_ug 1 5.86352686e+05 8.0177e+04 13.6739%
          P0_uxu_uux 2 3.49749021e+05 7.1386e+04 20.4107%
            P0_gg_gg 4 -2.64929351e+05 6.9954e+04 -26.4049%
           P0_gg_uux 1 7.28047092e+05 6.4983e+04 8.9257%
          P0_uxg_uxg 1 2.54532072e+05 4.8686e+04 19.1276%
          P0_gux_uxg 1 2.25260621e+05 4.7659e+04 21.1573%
        P0_uxux_uxux 1 1.42366408e+04 4.4670e+04 313.7646%
           P0_uux_gg 2 9.08958166e+05 2.2709e+04 2.4983%
           P0_uxu_gg 2 8.95828516e+05 1.8465e+04 2.0612%
          P0_uux_ccx 1 1.03301438e+05 3.0224e+03 2.9258%

ABS cross-section per integration channel:
            P0_gg_gg 5 1.76806907e+09 1.8657e+07 1.0552%
            P0_gu_ug 3 2.01904982e+08 6.5692e+06 3.2536%
            P0_ug_ug 2 2.04579989e+08 4.9852e+06 2.4368%
          P0_gux_uxg 3 1.15071537e+08 3.3744e+06 2.9324%
          P0_uxg_uxg 2 1.12394032e+08 2.8630e+06 2.5473%
            P0_gu_ug 2 1.21993528e+07 4.1156e+05 3.3736%
           P0_gg_uux 3 1.96937115e+07 3.9438e+05 2.0026%
            P0_uc_uc 1 1.36132886e+07 3.8470e+05 2.8259%
           P0_gg_uux 2 1.86550066e+07 3.5473e+05 1.9015%
            P0_ug_ug 3 1.22306963e+07 3.0048e+05 2.4567%
          P0_ucx_ucx 1 9.35955190e+06 2.3741e+05 2.5366%
          P0_uxc_uxc 1 9.32656293e+06 2.0638e+05 2.2128%
            P0_uu_uu 1 5.36101030e+06 2.0009e+05 3.7323%
          P0_gux_uxg 2 7.49159143e+06 1.6014e+05 2.1376%
        P0_uxcx_uxcx 1 5.90962458e+06 1.4054e+05 2.3781%
          P0_uxg_uxg 3 7.25625312e+06 1.1034e+05 1.5206%
          P0_uux_uux 2 3.43445190e+06 9.2671e+04 2.6983%
            P0_gu_ug 1 3.91764036e+06 8.8196e+04 2.2512%
            P0_ug_ug 1 3.76872489e+06 7.7462e+04 2.0554%
          P0_uxu_uux 2 3.50389116e+06 7.4969e+04 2.1396%
            P0_gg_gg 4 2.96679687e+06 6.9830e+04 2.3537%
           P0_gg_uux 1 4.23114396e+06 6.3549e+04 1.5019%
          P0_uxg_uxg 1 2.33727950e+06 4.7720e+04 2.0417%
          P0_gux_uxg 1 2.35719771e+06 4.6570e+04 1.9757%
        P0_uxux_uxux 1 1.95158343e+06 4.5474e+04 2.3301%
           P0_uux_gg 2 1.17129318e+06 2.6806e+04 2.2886%
           P0_uxu_gg 2 1.12773879e+06 1.9408e+04 1.7210%
          P0_uux_ccx 1 2.11287650e+05 2.9388e+03 1.3909%

Total ABS and
Total:
                      2.55409529e+09 +- 2.0895e+07 (8.1811e-01%)
                      -1.69475682e+08 +- 2.0529e+07 (-1.2114e+01%)

------------------------------------------------------------------------------------------------------------------------------------------------

This cross-section is 20% of the LO one in magnitude. As far as I understand, the [QCD] tag should include LO + NLO, not just NLO. Do you have any idea what could be causing this negative total cross-section?

Many thanks and kind regards,
Matthew De Angelis

Question information

Language:
English Edit question
Status:
Answered
For:
MadGraph5_aMC@NLO Edit question
Assignee:
Hua-Sheng Shao Edit question
Last query:
Last reply:
Revision history for this message
Hua-Sheng Shao (erdissshaw) said : 		#1

Hi Matthew De Angelis,

From your message, it looks you want to run fixed-order only, however you are running NLO+PS. With your setup, I tried it and I reproduced the negative total cross section.

The problem comes from the PDF you are using. I see you used 'cteq6_l' which is a LO PDF with your NLO calculations. The using of this LO PDF result in negative cross section. I tried the default NLO PDF 'nn23nlo'. The cross section looks reasonable. Can you tell me why you want to use LO PDF here ?

Cheers,

Hua-Sheng

Revision history for this message
Hua-Sheng Shao (erdissshaw) said : 		#2

Hi Matthew De Angelis,

Another important point I forgot to mention is: in dijet process at fixed order of NLO, you cannot apply the same jet pt cut because of the bording effect in counterevent. It would make the number to be reasonable only when you applied different jet cuts for the two jets. For example, in SubProcesses/cuts.f, I added the lines

c Apply the jet cuts
         if(DSQRT(pjet(1,1)**2+pjet(2,1)**2).LT.sycut+10d0)then
            passcuts_user=.false.
            return
         endif

at line 254,i.e. the leading jet pt cut is always 10 GeV bigger than the general jet pt cut in run_card.dat.

Cheers,

Hua-Sheng

Revision history for this message
Matthew De Angelis (matthew-deangelis) said : 		#3

Dear Hua-Sheng Shao,

Thank you for your prompt reply. I would like to run fixed-order only. I used the command 'launch DIJET -p' to get event files which I wish to run through a rivet analysis. Whilst the summary file says 'Parton shower to be used: HERWIG6', MG5_aMC posted a message after the generation process was complete stating that no showering has occurred and these events are unphysical. As for the PDF, I don't wish to use a LO PDF. I need to use the CTEQ6 NLO PDF, which is what I thought I was using - I got the cteq6_l from the minimal user guide V4.1 (https://cp3.irmp.ucl.ac.be/projects/madgraph/attachment/wiki/ManualAndHelp/Manual-March-2007.pdf) (P.7). Would you know what PDF name I require? Or alternatively, should I just call the PDF from LHAPDF? I have also managed to acquire a more reasonable total cross-section using the default NLO PDF 'nn23nlo' and the differing jet cuts you mentioned. I don't understand what is meant by boring effect in counterevent and the reason for differing jet cuts, would you be able to elaborate?

Kind regards,
Matthew

Revision history for this message
Hua-Sheng Shao (erdissshaw) said : 		#4

Yes, you can use LHAPDF. For the issue with same jet pt cut at fNLO, I assigned Marco. May be he can give a better explaination to you.

Revision history for this message
marco zaro (marco-zaro) said : 		#5

Hi,
so, let me clarify some things:
the cross section for dijet production with symmetric cuts is not well defined at NLO. You need to either ask for asymmetric cuts (i.e. two different pt cuts for the hardest and second jet), or to ask for some other kind of cuts, e.g. cutting on the sum of the pts of all partons. This is true regardless of the PDF you use.

This said, you can either run at fixed order or at NLO+PS.
The problem of running at NLO+PS is that you need to generate events with looser cuts than what you require at the analysis level, because of the effect of the shower. However, for dijet the behaviour of the cross section with the pt is very steep, which means that this kind of generation will be cursed by huge inefficiencies (ie. you need to generate huge event samples, ~ 10M events, to have a decent statistics after cuts). Note that you cannot plot the parton-level LHE file that you get before showering, as the events are not physical unless showered.

if you run at fixed order, you cannot generate events, rather you need to fill plots on the fly.
This is done by the Fixed order analyses, (you find some examples in the FixedOrderAnalysis directory, and you can choose which one to use with the FO_analyse_card. However, this way you cannot (to my knowledge) feed it directly to RIVET, rather you directly produce some histograms.

this is a bit the status.

Finally, you can for sure link lhapdf (both v6 or v5, only v5 supports the old cteq sets), just edit the run_card and set
lhapdf = pdlabel
and the corresponding LHAGLUE number for lhaid (it should be 10000 for cteq6m).
Finally you need to put the path to lhapdf-config (inside the bin dir of your lhapdf installation) inside
Cards/amcatnlo_configuration.txt,
for the lhapdf variable...

that should be all, i think...
cheers.

Marco

Can you help with this problem?

Provide an answer of your own, or ask Matthew De Angelis for more information if necessary.

To post a message you must log in.