PARTICLE_CONTAINER and Comix

[Frank Siegert]

The keyword 'PARTICLE_CONTAINER' allows to define custom particle containers, which is useful to simplify process listings. In some cases though, these containers will show unexpected behaviour compared to using their contents directly.

Example Run.dat without using the containers integrates fine:

(run){
  BEAM_1=2212
  BEAM_2=2212
  BEAM_ENERGY_1=3500
  BEAM_ENERGY_2=3500

  MASSIVE[5]=1
}(run)

(processes){
  Process 1 -1 -> 5 -5
  End process;
}(processes)

(selector){
  PT 5 20.0 E_CMS
}(selector)

Example Run.dat with containers integrates to nan:

(run){
  BEAM_1=2212
  BEAM_2=2212
  BEAM_ENERGY_1=3500
  BEAM_ENERGY_2=3500

  MASSIVE[5]=1
  PARTICLE_CONTAINER 905 bottoms 5 -5
}(run)

(processes){
  Process 1 -1 -> 905 905
  End process;
}(processes)

(selector){
  PT 905 20.0 E_CMS
}(selector)

Running this with Comix, the Amplitude::EvaluateAll returns a nan even though the momenta look reasonable (large enough pT).

[Frank Siegert]

I have simplified the Run.dat to leave out the METS selector and show that the same issue happens even with just a normal pT selector. Also, the nan seems to be Comix specific... at least a simple process like in the modified Run.dat's above does work with Amegic (which seems to have a different issue with containers, filed in a separate ticket). Running this with Comix, the Amplitude::EvaluateAll returns a nan even though the momenta look reasonable (large enough pT).

[Stefan Hoeche]

Fixed on trunk r18737.