ScaleSetter warnings: No CSS history in VBF-like process

[Frank Siegert]

The following run card for VBF like processes:

(run){
  BEAM_1=2212; BEAM_ENERGY_1=3500;
  BEAM_2=2212; BEAM_ENERGY_2=3500;
  ME_SIGNAL_GENERATOR   = Comix
}(run)

(processes){
  Process : 93 93 ->  93 93 25[a]
  Decay : 25[a] -> -24[b] 24[c]
  Decay : -24[b] -> 11 -12
  Decay : 24[c] -> -13 14
  Order_QCD : 0
  Integration_Error : 0.02
  End process
}(processes)

(selector){
  NJetFinder  2  15.0  0.0  0.4  -1
}(selector)

leads during integration (after the first iteration) to the following warnings for basically all processes:

METS_Scale_Setter::CalculateMyScale(): No CSS history for '2_6__d__db__h0[W-[e-__nu_eb]__W+[nu_e__e+]]__s__sb' in >35.55% of calls. Set \hat{s}.

[Frank Siegert]

I have just noticed that this also gives the occasional four-momentum violation when generating events.

[Frank Siegert]

The scale setter issue is still present with trunk r16943 and shows up in the benchmark PL_HVBF setup.

[Frank Siegert]

And possibly related: WW+jets shows METS_Scale_Setter::SetScales(): Failed to determine \mu. during integration in the benchmark setup.

[Stefan Hoeche]

The 'CSS history' warnings do not indicate a problem. They occcur as a result of invalid colour configurations, eventually leading to zero differential cross sections. As the scale setter is invoked before the ME calculation, there is no chance to avoid these messages at present.

[Stefan Hoeche]

The errors in WW+jets should be fixed with r16965.

[Stefan Hoeche]

Milestone rel-1.3.0 deleted