MadGraph5_aMC@NLO

Error in Weizsäcker-Williams implementation

Asked by Sebastian Urrutia-Quiroga

Hi everyone,
I am reproducing the results in Ref. [https://arxiv.org/abs/1105.0308], where they simulate pγ → ℓ+jjjν. I have learned that the IWW approximation is available in MG5, so I proceeded to run it emulating the syntax in Ref. [https://arxiv.org/abs/1602.06957]:

generate a p > e+ ve j j j
add process p a > e+ ve j j j
output IWW
lauch

I encountered some issues when setting lpp1 (lpp2) and pdlabel1 (pdlabel2):

analysis=off
set lpp1 3
set lpp2 1
set pdlabel1 iww
set pdlabel2 lhapdf

**************
Error when reading /Users/surrutiaquir/Software/MG5_aMC_v3_6_2/IWW/SubProcesses/P2_ga_lvlgqq/G1/results.dat
Command "generate_events run_01" interrupted with error:
Exception : Reported error: End code 23.0
  Full associated log:
  Process in group number 2
  A PDF is used, so alpha_s(MZ) is going to be modified
  Old value of alpha_s from param_card: 0.13000000000000003
  New value of alpha_s from PDF mixed : 0.11799999999999999
  pdf mixed not implemented in get_pdfup.
  known pdfs are
  none eva iww edff chff dressedmrs02nlmrs02nncteq4_mcteq4_lcteq4_dcteq5_mcteq5_dcteq5_lcteq5m1cteq6_mcteq6_lcteq6l1nn23lo nn23lo1nn23nlo
  using cteq5_m
  Define smin to 3700.0000000000000
  *****************************************************
  * MadGraph/MadEvent *
  * -------------------------------- *
  * http://madgraph.hep.uiuc.edu *
  * http://madgraph.phys.ucl.ac.be *
  * http://madgraph.roma2.infn.it *
  * -------------------------------- *
  * *
  * PARAMETER AND COUPLING VALUES *
  * *
  *****************************************************

   External Params
   ---------------------------------

  mdl_MB = 4.7000000000000002
  mdl_MT = 173.00000000000000
  mdl_MTA = 1.7769999999999999
  mdl_MZ = 91.188000000000002
  mdl_MH = 125.00000000000000
  aEWM1 = 132.50700000000001
  mdl_Gf = 1.1663900000000000E-005
  aS = 0.13000000000000003
  mdl_ymb = 4.7000000000000002
  mdl_ymt = 173.00000000000000
  mdl_ymtau = 1.7769999999999999
  mdl_WT = 1.4915000000000000
  mdl_WZ = 2.4414039999999999
  mdl_WW = 2.0476000000000001
  mdl_WH = 6.3823389999999999E-003
   Internal Params
   ---------------------------------

  mdl_CKM3x3 = 1.0000000000000000
  mdl_conjg__CKM1x1 = 1.0000000000000000
  mdl_conjg__CKM3x3 = 1.0000000000000000
  mdl_complexi = (0.0000000000000000,1.0000000000000000)
  mdl_MZ__exp__2 = 8315.2513440000002
  mdl_MZ__exp__4 = 69143404.913893804
  mdl_sqrt__2 = 1.4142135623730951
  mdl_MH__exp__2 = 15625.000000000000
  mdl_aEW = 7.5467711139788835E-003
  mdl_MW = 80.419002445756163
  mdl_sqrt__aEW = 8.6872153846781555E-002
  mdl_ee = 0.30795376724436879
  mdl_MW__exp__2 = 6467.2159543705357
  mdl_sw2 = 0.22224648578577766
  mdl_cw = 0.88190334743339216
  mdl_sqrt__sw2 = 0.47143025548407230
  mdl_sw = 0.47143025548407230
  mdl_g1 = 0.34919219678733299
  mdl_gw = 0.65323293034757990
  mdl_vev = 246.21845810181637
  mdl_vev__exp__2 = 60623.529110035903
  mdl_lam = 0.12886910601690263
  mdl_yb = 2.6995554250465490E-002
  mdl_yt = 0.99366614581500623
  mdl_ytau = 1.0206617000654717E-002
  mdl_muH = 88.388347648318430
  mdl_I1x33 = (2.69955542504654901E-002,0.0000000000000000)
  mdl_I2x33 = (0.99366614581500623,0.0000000000000000)
  mdl_I3x33 = (0.99366614581500623,0.0000000000000000)
  mdl_I4x33 = (2.69955542504654901E-002,0.0000000000000000)
  mdl_ee__exp__2 = 9.4835522759998875E-002
  mdl_sw__exp__2 = 0.22224648578577769
  mdl_cw__exp__2 = 0.77775351421422245
   Internal Params evaluated point by point
   ----------------------------------------

  mdl_sqrt__aS = 0.36055512754639896
  mdl_G__exp__2 = 1.6336281798666927
   Couplings of sm
   ---------------------------------

         GC_10 = -0.12781E+01 0.00000E+00
         GC_11 = 0.00000E+00 0.12781E+01
          GC_1 = -0.00000E+00 -0.10265E+00
          GC_2 = 0.00000E+00 0.20530E+00
          GC_3 = -0.00000E+00 -0.30795E+00
        GC_100 = 0.00000E+00 0.46191E+00

  Collider parameters:
  --------------------

  Running at e-P machine @ 1183.2159566199232 GeV
  PDF set = mixed
  alpha_s(Mz)= 0.1180 running at 2 loops.
  alpha_s(Mz)= 0.1180 running at 2 loops.
  Renormalization scale set on event-by-event basis
  Factorization scales fixed for beam1 @ 91.188000000000002 91.188000000000002

  getting user params
 Enter number of events and max and min iterations:
  Number of events and iterations 1000 5 3
 Enter desired fractional accuracy:
  Desired fractional accuracy: 0.10000000000000001
 Enter 0 for fixed, 2 for adjustable grid:
 Suppress amplitude (0 no, 1 yes)?
  Using suppressed amplitude.
 Exact helicity sum (0 yes, n = number/event)?
  Explicitly summing over helicities
 Enter Configuration Number:
 Running Configuration Number: 1
  Not subdividing B.W.
  Attempting mappinvarients 1 7
  Determine nb_t
  T-channel found: 4
  Completed mapping 7
  about to integrate 13 1000 7 3 13 1
  Using non-zero grid deformation.
  13 dimensions 1000 events 13 invarients 7 iterations 1 config(s), (0.99)
  Using h-tuple random number sequence.
  Error opening grid
  Using Uniform Grid! 28
  Using uniform alpha 1.0000000000000000
  Grid defined OK
  Masses: 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00 0.000E+00
  Set CM energy to 1183.22
  Mapping Graph 1 to config 1
  Determine nb_t
  T-channel found: 4
 Setting grid 1 0.71429E-04 1
 Setting grid 2 0.28571E-03 1
  Warning: No cutoff for shat integral found
           Minimum set to 1.0000000000000000E-004
 Setting grid 3 0.10000E-03 1
 Setting grid 4 0.28571E-03 1
  Transforming s_hat 1/s 12 3.4999978161936018E-003 3700.0000000000000 1400000.8735231042
 Using random seed offsets 1 : 3
   with seed 30
  Ranmar initialization seeds 8292 9407
   Particle 3 4 5 6 7
       Et > 10.0 0.0 20.0 20.0 20.0
        E > 0.0 0.0 0.0 0.0 0.0
      Eta < 2.5 -1.0 5.0 5.0 5.0
    xqcut: 0.0 0.0 0.0 0.0 0.0
 d R # 3 > -0.0 0.0 0.4 0.4 0.4
 d R # 4 > -0.0 -0.0 0.0 0.0 0.0
 d R # 5 > -0.0 -0.0 -0.0 0.4 0.4
 d R # 6 > -0.0 -0.0 -0.0 -0.0 0.4
 s min # 3> 0.0 0.0 0.0 0.0 0.0
 s min # 4> 0.0 0.0 0.0 0.0 0.0
 s min # 5> 0.0 0.0 0.0 0.0 0.0
 s min # 6> 0.0 0.0 0.0 0.0 0.0
 xqcutij # 3> 0.0 0.0 0.0 0.0 0.0
 xqcutij # 4> 0.0 0.0 0.0 0.0 0.0
 xqcutij # 5> 0.0 0.0 0.0 0.0 0.0
 xqcutij # 6> 0.0 0.0 0.0 0.0 0.0
  No cut BW -4 1

  ********************************************
  * You are using the DiscreteSampler module *
  * part of the MG5_aMC framework *
  * Author: Valentin Hirschi *
  ********************************************

  impossible call (or was it) to pdf-> please reporrt
 STOP 23

 ls status:
 input_app.txt
 log.txt
 run1_app.log

Please report this bug on https://bugs.launchpad.net/mg5amcnlo
More information is found in '/Users/surrutiaquir/Software/MG5_aMC_v3_6_2/IWW/run_01_tag_1_debug.log'.
Please attach this file to your report.
**************

I am probably doing something terribly wrong, so I would appreciate any guidance. Thanks!

Question information

Language:
English Edit question
Status:
Solved
For:
MadGraph5_aMC@NLO Edit question
Assignee:
No assignee Edit question
Solved by:
Olivier Mattelaer
Solved:
Last query:
Last reply:
Revision history for this message
Olivier Mattelaer (olivier-mattelaer) said : 		#1

I have tested the following code:

 generate a p > e+ ve j j j
 output
 launch
 set lpp1 2
 set lpp2 1
 set pdlabel1 iww
 set pdlabel2 lhapdf

and it works.

Note that
1. I do not include the symmetric process (p a) because otherwise the pdf can not be defined consistently
2. I use lpp1=2 and not lpp1=3

since as written in the cards:
  28 #*********************************************************************
  29 # Collider type and energy *
  30 # lpp: 0=No PDF, 1=proton, -1=antiproton, *
  31 # 2=elastic photon of proton/ion beam *
  32 # +/-3=PDF of electron/positron beam *
  33 # +/-4=PDF of muon/antimuon beam *
  34 #*********************************************************************

So "3" means that you have a electron/positron beam, while your question seems to indicate that you are looking for "2".

Cheers,

Olivier

Revision history for this message
Sebastian Urrutia-Quiroga (sgurruti) said : 		#2

Hi Olivier,

Thanks for your prompt response. If you don't mind, I have a few follow-up questions:

1) Ref. [https://arxiv.org/abs/1105.0308] mentions that "[...] PDF of a photon inside the charged lepton." That's why I presumed I should use lpp1=3.

2) In Ref. [https://arxiv.org/abs/1602.06957], the authors explicitly include the symmetric process, in a different context, though:
     > define q = u c d s b u~ c~ d~ s~ b~
     > generate q a > n2 mu q QED=3 QCD=0 [QCD]
     > add process a q > n2 mu q QED=3 QCD=0 [QCD]
     > output PP_VBF_NLO; launch;
Why would that be necessary?

Many thanks again!

Revision history for this message
Best Olivier Mattelaer (olivier-mattelaer) said : 		#3

To answer your question, one needs to know the physics that you want to do.
Since your syntax/question was ambiguous about that, I did assume that you were looking for proton-proton collision.

If you are looking for electron-proton collision, then yes it does make sense to use lpp1=3
On the other hand, if you are looking for electron-protron collision, then it makes no sense to add the symmetric process like you did since the assymetry is at the beam level. Maybe asking for something like
generate e- a > ...
add process a p >
can make sense but asking for
generate q a > ...
add process a q > ...
is close to impossible since you will ask the PDF of a quark inside the electron.
Now this is true that some PDF might provide such type of PDF which will then allow such type of computation.
But this is the line that make me think that you were looking for pp.

Independently, the lpp1 applies to the first beam and lpp2 on the second beam, so if lpp1 is different of lpp2, you will likely not be able to use the "add process" since the value lpp1 will then be wrong in one case.

Cheers,

Olivier

Revision history for this message
Sebastian Urrutia-Quiroga (sgurruti) said : 		#4

Thanks Olivier Mattelaer, that solved my question.

Revision history for this message
David Henry (davidhenry46) said : 		#5

It looks like the error is tied to the mixed PDF handling in MG5 — the message about “pdf mixed not implemented in get_pdfup” suggests the IWW interface isn’t fully integrated with certain PDF labels. You may want to try switching both beams to supported PDFs (like cteq5_m or nn23lo) and see if the run completes. Others have reported similar issues with the iww label on beam 1 when mixing with lhapdf.

While digging into this, I also like to keep side projects for a change of pace — for example, I’ve been working on https://birthdaywishfriends.com/
, which is a lighter creative space. Sometimes stepping away and returning with a fresh mindset makes debugging these long error logs a lot easier.

Revision history for this message
Wasif Ali (wasifali758595) said : 		#6

An error in the Weizsäcker-Williams (WW) implementation usually occurs when applying the equivalent photon approximation in particle physics simulations or calculations. The WW method simplifies complex electromagnetic interactions by treating fast-moving charged particles as a flux of virtual photons. Errors typically arise from:

Common Causes

Incorrect Kinematic Limits

Choosing inappropriate energy, momentum transfer, or angular ranges can violate the assumptions of the WW approximation.

Numerical Instabilities

Integrals over photon flux or cross-sections may produce errors if the step size is too large or the integration method is unstable.

Misinterpretation of the Photon Flux Formula

Using the wrong form factor or neglecting relevant particle structure effects can lead to overestimation or underestimation of rates.

Software/Code Implementation Issues

Many simulation frameworks (e.g., MadGraph, PYTHIA) provide WW implementations. Errors may arise from outdated versions, incorrect input parameters, or incompatible physics modules.

Recommended Solutions

Check Kinematic Assumptions

Verify that the energy and momentum ranges satisfy the WW approximation conditions.

Validate Form Factors

Ensure the proper charge distribution or nuclear form factor is used for the target particle.

Refine Numerical Methods

Use smaller integration steps or more robust numerical algorithms if instability is observed.

Compare with Benchmark Results

Cross-check outputs against analytical calculations or published results to identify inconsistencies.

Update Software/Modules

Ensure the simulation package or code implementation is the latest version and correctly configured.

In short: WW implementation errors usually stem from kinematic, numerical, or input issues. Careful verification of assumptions, parameters, and numerical methods is essential to resolve them.

Revision history for this message
Sebastian Urrutia-Quiroga (sgurruti) said : 		#7

AI/ChatGPT (not) to the rescue. Sorry about that to the entire community.

Revision history for this message
Gary Eric (infogaryeric) said : 		#8

Tropical Menu https://tropicalsmoothiemenus.com/ serves up paradise-inspired food and drinks, packed with juicy fruits, bright colors, and beachy vibes that taste like summer all year long.