Clustering fails with r17279

[Frank Siegert]

The cluster algorithm on trunk fails in the following process (Failed to determine configuration. Rejecting point.):

(run){
  EVENTS 1000000;
  OUTPUT 2;
  MASS[25]=165.
  YUKAWA_B=0.;
  FRAGMENTATION Off;

  BEAM_1  2212; BEAM_ENERGY_1 3500;
  BEAM_2  2212; BEAM_ENERGY_2 3500;

  ME_SIGNAL_GENERATOR Amegic;
  EVENT_GENERATION_MODE Weighted;
  SCALES VAR{MPerp2(p[2]+p[3]+p[4]+p[5])};

  MODEL=SM+EHC;
  FINITE_TOP_MASS=0;
}(run);

(processes){
  Process 93 93 -> 25[a] 93{1}
    Decay   25[a] -> 24[c] -24[d];
      Decay   24[c] -> -11  12;
      Decay  -24[d] ->  13 -14;
  CKKW sqr(30.0/E_CMS);
  Order_EW 1;
  End process;
}(processes);

(selector){
  Mass  -11  12 10. E_CMS
  Mass   13 -14 10. E_CMS
}(selector)

This does not happen without c17279.

With the default EXCLUSIVE_CLUSTER_MODE=0 it rejects ~60% of all points, while with EXCLUSIVE_CLUSTER_MODE=1 (i.e. only QCD clusterings) it rejects 100%.

Could this be related to another recent change, namely clustering the explicit decays first, which would in this case here lead to a 2->1 configuration? Or is it somehow related to the SM+EHC vertex being special?

[Frank Siegert]

Here is a slightly simpler process setup that shows the same symptoms (modified from #161):

(run){
  EVENTS 1000000;

  BEAM_1  2212; BEAM_ENERGY_1 3500;
  BEAM_2  2212; BEAM_ENERGY_2 3500;

  ME_SIGNAL_GENERATOR Amegic;
  SCALES VAR{Abs2(p[0]+p[1])};

  MODEL      = SM+U1_B
  M_Z' = 160;                                                                                                                                                                                               
  g'_1 = 0.5;
  MIX  = 0;
}(run);

(processes){
  Process 2 2 -> 6[a] 6[b]
  Decay  6[a] -> -13  14  5
  Decay  6[b] ->  93  93  5
  Order_EW  2;
  Order_QCD 0;
  End process;
}(processes);

(selector){
  Mass  13 -14 10 E_CMS;
  Mass -13  14 10 E_CMS;
  PT     5     10 E_CMS;
  PT    -5     10 E_CMS;
}(selector)

(integration){
  INTEGRATOR 7; ERROR 0.1;
  FINISH_OPTIMIZATION On;
}(integration);

Interesting to note is that both of these are non-SM setups, so maybe it's related to that? I have tested Amegic for the same process but with 2 -2 -> 6 -6 in the SM and it seems to work properly there.

[Stefan Hoeche]

This problem is solved with r17316 and r17317.

[Stefan Hoeche]

Milestone rel-1.3.0 deleted