Accessing Pythia6 Processes

Access PYTHIA 6 Processes

Gradually all relevant processes from PYTHIA 6 are being re-implemented in PYTHIA 8, but this transition is not finished. For a while longer it may therefore at times be convenient to access the Fortran PYTHIA 6 process library. In order to give this access at runtime, and not only by writing/reading event files, an interface is provided to C++. This interface is residing in Pythia6Interface.h, and in addition the PYTHIA 6 library must be linked. The latter should normally be the most recent Fortran PYTHIA version, but must be at least 6.314, since this is the first version that allows processes to be output in the Les Houches format (and not only input).

The routines interfaced are

pygive(command) to give in a command to change a setting, e.g. to decide which processes should be generated,
pyinit( frame, beam, target, wIn) to initialize the event generation chain,
pyupin() to fill this initialization information in the HEPRUP commonblock,
pyupev() to generate the next hard process and fill the event information in the HEPEUP commonblock,
pylist( mode) to list the event at the process level,
pystat( mode) to print statistics on the event generation process.

Details on allowed arguments are given in the PYTHIA 6.4 manual.

These methods can be used in context of the LHAupFortran class. The existing code there takes care of converting HEPRUP and HEPEUP commonblock information from Fortran to C++, and of making it available to the PYTHIA 8 methods. What needs to be supplied are the two LHAupFortran::fillHepRup() and LHAupFortran::fillHepEup() methods. The first can contain an arbitrary number of pygive(...), followed by pyinit(...) and pyupin() in that order. The second only needs to contain pyupev(). Finally, the use of pylist(...) and pystat(...) is entirely optional, and calls are best put directly in the main program.

This means that all other Fortran routines have not been interfaced and cannot be accessed directly from the C++ code; there is no need for them in the current setup.

PYTHIA 6.4 does its own rejection of events internally, according to the strategy option 3. However, the current runtime interface does not take cross-section information from PYTHIA 6.4. This means that both the initial maxima and the final cross sections printed by the PYTHIA 8 routines are irrelevant in this case. Instead you have to study the standard PYTHIA 6.4 initialization printout and call on pystat(...) at the end of the run to obtain this information. It also means that you cannot mix with internally generated events, unlike what could have been allowed for strategy 3. Should a strong need arise, PYTHIA 6.4 could be modified to work with strategy option 1 and thus allow a mixing with internal processes, but we do not expect this to happen.

An example of a fillHepRup() method to set up Z^0 production at LHC, with masses above 50 GeV, would be

bool LHAinitFortran::fillHepRup() { 
  Pythia6Interface::pygive("msel = 11"); 
  Pythia6Interface::pygive("ckin(1) = 50."); 
  Pythia6Interface::pyinit("cms","p","p",14000.);   
  Pythia6Interface::pyupin();
  return true;
}

and the process-generic fillHepEup() method would be

bool LHAupFortran::fillHepEup() { 
  Pythia6Interface::pyupev();
  return true;
}

Note that, of all parameters that could be set with the PYGIVE, only those that influence the generation of the hard processes have any impact, since this is the only part of the Fortran code that is used. Also, if you want to set many parameters, you can easily collect them in one file (separate from PYTHIA 8 input) and parse that file.

All hard PYTHIA 6 processes should be available for full generation in PYTHIA 8, at least to the extent that they are defined for beams of protons and antiprotons, which is the key application for PYTHIA 8 so far. Soft processes, i.e. elastic and diffractive scattering, as well as minimum-bias events, require a different kinematics machinery, and can only be generated with the internal PYTHIA 8 processes.

A simple example is found in main51.cc, another with parsed input in main52.cc and a third with HepMC output in main54.cc.