From EIC
Revision as of 13:50, 19 December 2011 by Elke (talk | contribs) (→‎Running PEPSI)
Jump to navigation Jump to search

PEPSI (Polarised Electron Proton Scattering Interactions) [1] is a Monte Carlo generator for polarised deep inelastic scattering (pDIS). It is based on the LEPTO 6.5 Monte Carlo for unpolarised DIS.

PEPSI References

(1) "PEPSI - a Monte Carlo generator for polarized leptoproduction", L. Mankiewicz, A. Schäfer and M. Veltri, Comp. Phys. Comm. 71, 305-318 (1992).

Parton distribution functions

The distribution function to use in polarised leptoproduction is set via the variable LST(15) in the LEPTO COMMON block /LEPTOU/. Tables 1 and 2 list internal allowed values of LST(15) for polarised and unpolarised distributions respectively.

Pepsi is linked with the pdflib such all PDFs included in there can be used by setting LST(15) to the respective PDF-ID

PEPSI processes important in ep

Subprocess # Description
γ∗q → q 1 LO DIS
γ∗ q → qg 2 QCDC
γ∗ g → q qbar 3 PGF

QCDC: QCD-Compton, radiation of a gluon from incoming or outgoing quark lines
PGF: Photon Gluon Fusion

Running PEPSI

the code can be found on the afs directory for EIC at BNL


this code is based on PEPSI/LEPTO and was modified to include radiative corrections using RadGen.
The main program is in the same directory and called pepsiMaineRHIC_noradcorr.f / pepsiMaineRHIC_radcorr.f, several other routines are needed, which are in the same directory.
The executable is in the same directory and called pepsieRHICnoRAD / pepsieRHICwithRAD
There are several steer files (named: XXXXX.eic) provided in this directory to run PEPSI and get reasonable output.

How to Run the Code


PEPSI has to be run twice if polarized asymmetries should be generated, once for parallel, lepton and proton beam spin direction parallel, and antiparallel spin state.

without radiative corrections

 pepsieRHICwithRAD < input.data_noradcor.eic.pol.anti > XXX.log  

input.data_noradcor.eic.pol.anti is one of the steer file examples in the directory to run PEPSI with settings tuned for Hermes, and/or H1 and ZEUS for the antiparallel polarized cross-section
input.data_noradcor.eic.pol.par: for the parallel polarized cross-section
input.data_noradcor.eic.unpol: for the unpolarised cross-section

with radiative corrections

  • create a directory called radgen in the area you want to run the code
  • you either need to generate the lookup table for your cuts and beam energy settings first
   pepsieRHICwithRAD < input.data_make-radcor.eic.unpol
  • or you can use one of the files already generated

the files are in the directory


and called xytab1unp.04.050.dat or xytab1ant.04.050.dat or xytab1par.04.050.dat

  • to run the code than with radiative corrections simply change the steer file to either

input.data_radcor.eic.unpol or input.data_radcor.eic.pol.par or .... and

 type pepsieRHICwithRAD < input.data_radcor.eic.unpol > XXX.log

Output file structure

the output file is in a text format which has the following structure.

  • 1st line: PEPSI EVENT FILE
  • 2nd line: "============================================"

  • 3rd line: Information on event wise variables stored in the file
I: 0 (line index)
ievent: eventnumber running from 1 to XXX
genevent: trials to generate this event
process: pepsi subprocess (LST(23)), for details see table above
subprocess: pythia subprocess (LST(24)), for details see table above
nucleon: hadron beam type (LST(22))
struckparton: parton hit in the target (LST(25))
partontrack: # or parton track (LST(26))
truey, trueQ2, truex, trueW2, trueNu: are the kinematic variables of the event.
If radiative corrections are turned on they are different from what is calculated from the scattered lepton.
If radiative corrections are turned off they are the same as what is calculated from the scattered lepton
mcfixedweight weight calculated from generation limits
weight total weight including everything
dxsec cross section included in the weight
mcextraweight Pepsi total cross section in pb from numerical integration parl(23)
dilut, F1, F2, A1, A2, R, Depol, d, eta, eps, chi true variables needed to calculate g1
gendilut, genF1, genF2, genA1, genA2, genR, genDepol, gend, geneta, geneps, genchi variables needed to calculate g1
SigRadCor: information used and needed in the radiative correction code
EBrems: energy of the radiative photon in the nuclear rest frame
nrTracks: number of tracks in this event, includes also virtual particles

  • 4th line: "============================================"

  • 5th line: Information on track wise variables stored in the file
I: line index, runs from 1 to nrTracks
K(I,1): status code KS (1: stable particles 11: particles which decay 55; radiative photon)
K(I,2): particle KF code (211: pion, 2112:n, ....)
K(I,3): line number of parent particle
K(I,4): normally the line number of the first daughter; it is 0 for an undecayed particle or unfragmented parton
K(I,5): normally the line number of the last daughter; it is 0 for an undecayed particle or unfragmented parton.
P(I,1): px of particle
P(I,2): py of particle
P(I,3): pz of particle
P(I,4): Energy of particle
P(I,5): mass of particle
V(I,1): x vertex information
V(I,2): y vertex information
V(I,3): z vertex information
  • 6th line: "============================================"
  • 7th line: event information for first event
  • 8th line: "============================================"
  • 9th to X-1 line: trackwise info of 1st event
  • Xth line "=============== Event finished ==============="

the information from line 7 to X repeats for each event.

How to analyze events

  • create a root tree

there are root macros available to convert the output txt-files into root trees.
Details how to run the macros can be found [here]

MC normalization (still under construction)

to normalize your counts to cross section you need two informations

  • the total number of trials (NGEN(0,3)), it is printed to the screen/logfile if PYTHIA finishes
  • the total integrated cross section (PARI(1)), the unit is microbarn (10^-6), it is printed to the screen/logfile if PYTHIA finishes
==> count * total integrated cross section /total number of trials

to calculate the corresponding luminosity

==> total number of trials/ total integrated cross section

Documentation on Radiative Corrections:

the code implemented in PTHIA to calculate radiative corrections is called RADGEN
The writeup on it can be found here [hep-ph/9906408]
The following steps have been done to implement it in PYTHIA:

  • change the subroutine pygaga.f so it calls radgen after you have thrown y and Q2
  • get the true y and Q2 from radgen and the radiated photon
  • Pythia will continue to now generate an event based on this y and Q2
  • Pythia still operates under accept reject, the extra weigt from the radiative corrections is absorbed in the flux factor