Skip to content

ThOr migration of D -> hhh

Felipe Luan Souza De Almeida requested to merge felipe-d2hhh into charm-thor-hlt2-lines

Description

Migration of d2hhh hlt2 lines to use ThOr functors and common builders, following !1221 (merged).

Rates

Click to expand

Rates computed with upgrade_minbias_hlt1_filtered sample.

HLT rates:
----------------------------------------------------------------------------------------------------
Line:    Hlt2CharmHadDToKmKpKpLineDecision       Incl: 0.73 +/- 0.22 kHz, Excl: 0.13 +/- 0.09 kHz
Line:    Hlt2CharmHadDToKmKpPipLineDecision      Incl: 1.87 +/- 0.35 kHz, Excl: 0.80 +/- 0.23 kHz
Line:    Hlt2CharmHadDToKmPipPipLineDecision     Incl: 7.14 +/- 0.68 kHz, Excl: 5.27 +/- 0.59 kHz
Line:    Hlt2CharmHadDToKpKpPimLineDecision      Incl: 0.80 +/- 0.23 kHz, Excl: 0.33 +/- 0.14 kHz
Line:    Hlt2CharmHadDToKpPimPipLineDecision     Incl: 2.60 +/- 0.41 kHz, Excl: 1.07 +/- 0.26 kHz
Line:    Hlt2CharmHadDToPimPipPipLineDecision    Incl: 2.07 +/- 0.37 kHz, Excl: 1.20 +/- 0.28 kHz
Line:    Hlt2CharmHadDsToKmKpKpLineDecision      Incl: 1.40 +/- 0.30 kHz, Excl: 0.67 +/- 0.21 kHz
Line:    Hlt2CharmHadDsToKmKpPipLineDecision     Incl: 4.54 +/- 0.54 kHz, Excl: 2.47 +/- 0.40 kHz
Line:    Hlt2CharmHadDsToKpKpPimLineDecision     Incl: 2.00 +/- 0.36 kHz, Excl: 1.54 +/- 0.31 kHz
Line:    Hlt2CharmHadDsToKpPimPipLineDecision    Incl: 6.54 +/- 0.65 kHz, Excl: 4.27 +/- 0.53 kHz
Line:    Hlt2CharmHadDsToPimPipPipLineDecision   Incl: 1.40 +/- 0.30 kHz, Excl: 0.73 +/- 0.22 kHz
Line:    Hlt2CharmhadDsToKmPipPipLineDecision    Incl: 3.61 +/- 0.48 kHz, Excl: 2.14 +/- 0.37 kHz
----------------------------------------------------------------------------------------------------
Hlt2 Total:                                       Rate:  27 +/- 1.3 kHz

Rates_Hlt2

Efficiencies

Click to expand

Efficiencies computed with MC samples from sim+std://MC/Upgrade/Beam7000GeV-Upgrade-MagDown-Nu7.6-25ns-Pythia8/Sim10-Up05/Reco-Up03/. No MC sample was generated for the forbidden modes (Hlt2CharmHadDToKpKpPimLine and Hlt2CharmhadDsToKmPipPipLine). Integrated HLT efficiencies for the lines with denominator: CanRecoChildren.

  • D^{+} \rightarrow K^-K^+\pi^+: Hlt2CharmHadDToKmKpPipLine

Event type: 21163002, dkfile: D+_K-K+pi+=phsp,TightCut2

Line:    D_Hlt2CharmHadDToKmKpPipLineDecisionTOS         Efficiency:     0.229 +/- 0.014
Line:    Hlt2CharmHadDToKmKpPipLineDecision              Efficiency:     0.231 +/- 0.015
  • D_{s}^{+} \rightarrow K^-K^+\pi^+: Hlt2CharmHadDsToKmKpPipLine

Event type: 23163003, dkfile: Ds+_K-K+pi+=phsp,TightCut2

Line:    Ds_Hlt2CharmHadDsToKmKpPipLineDecisionTOS       Efficiency:     0.175 +/- 0.013
Line:    Hlt2CharmHadDsToKmKpPipLineDecision             Efficiency:     0.181 +/- 0.013
  • D^{+} \rightarrow K^-K^+K^+: Hlt2CharmHadDToKmKpKpLine

Event type: 21163042, dkfile: D+_K-K+K+=phsp,TightCut2

Line:    D_Hlt2CharmHadDToKmKpKpLineDecisionTOS  Efficiency:     0.196 +/- 0.013
Line:    Hlt2CharmHadDToKmKpKpLineDecision       Efficiency:     0.198 +/- 0.013
  • D_{s}^{+} \rightarrow K^-K^+K^+: Hlt2CharmHadDsToKmKpKpLine

Event type: 23103042, dkfile: Ds+_K-K+K+=phsp,TightCut2

Line:    Ds_Hlt2CharmHadDsToKmKpKpLineDecisionTOS        Efficiency:     0.164 +/- 0.013
Line:    Hlt2CharmHadDsToKmKpKpLineDecision              Efficiency:     0.167 +/- 0.013
  • D^{+} \rightarrow \pi^-\pi^+\pi^+: Hlt2CharmHadDToPimPipPipLine

Event type: 21163012, dkfile: D+_pi-pi+pi+=phsp,TightCut2

Line:    D_Hlt2CharmHadDToPimPipPipLineDecisionTOS       Efficiency:     0.249 +/- 0.013
Line:    Hlt2CharmHadDToPimPipPipLineDecision            Efficiency:     0.253 +/- 0.013
  • D_{s}^{+} \rightarrow \pi^-\pi^+\pi^+: Hlt2CharmHadDsToPimPipPipLine

Event type: 23103012, dkfile: Ds+_pi-pi+pi+=phsp,TightCut2

Line:    Ds_Hlt2CharmHadDsToPimPipPipLineDecisionTOS     Efficiency:     0.159 +/- 0.012
Line:    Hlt2CharmHadDsToPimPipPipLineDecision           Efficiency:     0.161 +/- 0.012
  • D^{+} \rightarrow K^-\pi^+\pi^+: Hlt2CharmHadDToKmPipPipLine

Event type: 21163022, dkfile: D+_K-pi+pi+=phsp,TightCut2

Line:    D_Hlt2CharmHadDToKmPipPipLineDecisionTOS        Efficiency:     0.250 +/- 0.015
Line:    Hlt2CharmHadDToKmPipPipLineDecision             Efficiency:     0.252 +/- 0.015
  • D^{+} \rightarrow \pi^-\pi^+K^+: Hlt2CharmHadDToKpPimPipLine

Event type: 21163032, dkfile: D+_pi-pi+K+=phsp,TightCut2

Line:    D_Hlt2CharmHadDToKpPimPipLineDecisionTOS        Efficiency:     0.201 +/- 0.013
Line:    Hlt2CharmHadDToKpPimPipLineDecision             Efficiency:     0.206 +/- 0.013
  • D_{s}^{+} \rightarrow \pi^-\pi^+K^+: Hlt2CharmHadDsToKmPipPipLine

Event type: 23103062, dkfile: Ds+_pi-pi+K+=phsp,TightCut2

Line:    Ds_Hlt2CharmHadDsToKpPimPipLineDecisionTOS      Efficiency:     0.187 +/- 0.013
Line:    Hlt2CharmHadDsToKpPimPipLineDecision            Efficiency:     0.190 +/- 0.013
  • D_{s}^{+} \rightarrow \pi^-K^+K^+: Hlt2CharmHadDsToKpKpPimLine

Event type: 23163052, dkfile: Ds+_pi-K+K+=phsp,TightCut2

Line:    Ds_Hlt2CharmHadDsToKpKpPimLineDecisionTOS       Efficiency:     0.173 +/- 0.013
Line:    Hlt2CharmHadDsToKpKpPimLineDecision             Efficiency:     0.177 +/- 0.013

Summary table

Line Incl. Rate (kHz) Excl. Rate (kHz) Eff. Hlt2 (%) True Rate (kHz) Expected Rate (kHz) *
D^{+} \rightarrow K^-K^+K^+ 0.73 \pm 0.22 0.13 \pm 0.09 19.8 \pm 1.3 0.10 0.0198 \pm 0.0013
D^{+} \rightarrow K^-K^+\pi^+ 1.87 \pm 0.35 0.80 \pm 0.23 23.1 \pm 1.5 16.22 3.75 \pm 0.24
D^{+} \rightarrow K^-\pi^+\pi^+ 7.14 \pm 0.68 5.27 \pm 0.59 25.2 \pm 1.5 157.21 39.6 \pm 2.4
D^{+} \rightarrow \pi^-K^+K^+ 0.80 \pm 0.23 0.33 \pm 0.14
D^{+} \rightarrow \pi^-\pi^+K^+ 2.60 \pm 0.41 1.07 \pm 0.26 20.6 \pm 1.3 0.82 0.169 \pm 0.011
D^{+} \rightarrow \pi^-\pi^+\pi^+ 2.07 \pm 0.37 1.20 \pm 0.28 25.3 \pm 1.3 5.48 1.386 \pm 0.071
D_{s}^{+} \rightarrow K^-K^+K^+ 1.40 \pm 0.30 0.67 \pm 0.21 16.7 \pm 1.3 0.15 0.0251 \pm 0.0020
D_{s}^{+} \rightarrow K^-K^+\pi^+ 4.54 \pm 0.54 2.47 \pm 0.40 18.1 \pm 1.3 38.05 6.89 \pm 0.49
D_{s}^{+} \rightarrow \pi^-K^+K^+ 2.00 \pm 0.36 1.54 $\pm$ 0.31 17.7 $\pm$ 1.3 0.09 0.0159 $\pm$ 0.0011
$D_{s}^{+} \rightarrow \pi^-\pi^+K^+$ 6.54 $\pm$ 0.65 4.27 $\pm$ 0.53 19.0 $\pm$ 1.3 4.59 0.872 $\pm$ 0.060
$D_{s}^{+} \rightarrow \pi^-\pi^+\pi^+$ 1.40 $\pm$ 0.30 0.73 $\pm$ 0.22 16.1 $\pm$ 1.2 7.62 1.227 $\pm$ 0.091
$D_{s}^{+} \rightarrow K^-\pi^+\pi^+$ 3.61 $\pm$ 0.48 2.14 $\pm$ 0.37

(*) Expected Rate computed as (True Rate) x (Eff. Hlt2)

Open Questions

  • We have very high rate modes. Should we add a prescale in these cases?
    • Marian: To be discussed

Closed Questions

Click to expand
  • Should we add (when the algorithm is implemented) a TMVA line with ThOr functors for some modes?
    • Marian: possibly once such a selection can be well understood. This will need monitoring efforts and a calibration strategy.
  • Do we need a ThOr functor for BPVALID()?
    • Marian: the prefilters include a filter on the existence of PVs in the event. In principle this is enough.
  • Currently we use a wrapper (_make_dplus()) to the common builder combine_3body() with cut values as arguments, since the selection is essentially the same for the different lines. This could be seen as error prone, since we might want to change some cut values for specific lines, but might end up affecting all lines at once. Should we remove the wrapper and use directly the combine_3body() call with the cut values for each line?
    • Our comment: in the current setup, the change of some cut value for a specific line can be achieved locally by passing the new cut value in the _make_dplus() call. For example, the current default cut value set for trk_2of3_ipchi2_min is trk_2of3_ipchi2_min=10, but we apply a different cut for the line Hlt2CharmHadDToKmKpKpLine. This is achieved simply by adding trk_2of3_ipchi2_min=4 as argument to the _make_dplus() call in this line configuration. So, to change the cut values for specific modes it is enough to add the different cut values to the wrapper call in the line configuration, which is safe in the sense it won't affect other lines, and we still benefit from the common builders, since in this way we do not interact directly with ThOr functors and there are useful runtime errors configured there. In any case, we remain open to change the implementation if requested.
    • Marian: as we cannot exclude that you are the only ones who edit this file, so I would recommend to declare cut values at a single place, and not overwrite them later.
    • Follow-up: wrappers removed

FYI: @amato @polye

Edited by Tommaso Pajero

Merge request reports