TopSanityCheckerValidation
Jump to navigation
Jump to search
CMSSW_3_1_X MC validation
- Environment settings :
cmsrel CMSSW_3_1_1 cd CMSSW_3_1_1/src cvs co -r V07-07-03 PhysicsTools/PatAlgos cvs co -r V06-02-02 AnalysisDataFormats/TopObjects cvs co -r V06-02-02 TopQuarkAnalysis cvs up -r V06-02-02-02 TopQuarkAnalysis/TopJetCombination mkdir TopBrussels; cd TopBrussels; cp -r /user/ghammad/CMSSW_3_1_1/src/TopBrussels/SanityChecker . cd .. ; scramv1 b -j 4;
- PAT-tuples :
- cfg file to be used :
- tt+jets : /user_mnt/user/ghammad/CMSSW_3_1_1/src/PhysicsTools/PatAlgos/test/mypatLayer1TtGenEvt_fromAOD_full_cfg.py
- others : /user_mnt/user/ghammad/CMSSW_3_1_1/src/PhysicsTools/PatAlgos/test/mypatLayer1_fromAOD_full_cfg.py
- location :
- WW : /pnfs/iihe/cms/store/user/ghammad/CMSSW31X/PreProd/Validation/WW
- InclusiveMu15 : /pnfs/iihe/cms/store/user/ghammad/CMSSW31X/PreProd/Validation/InclusiveMu15
- TTjets : /pnfs/iihe/cms/store/user/ghammad/CMSSW31X/PreProd/Validation/TTbar
- cfg file to be used :
- Sanity Checker :
- cfg file to be used :
- /user_mnt/user/ghammad/CMSSW_3_1_1/src/TopBrussels/SanityChecker/test/sanitychecker_cfg.py
- warning : dedicated "tt+jets" modules need to be commented if run over "non tt+jets" samples (more info in the cfg files)
- log files :
- tt+jets : /user_mnt/user/ghammad/CMSSW_3_1_1/src/TopBrussels/SanityChecker/test/TTbar
- comparison with Summer08 :
- CMSSW_3_1_1 tt+jets samples available (/TTbar/Summer09-MC_31X_V2_preproduction_311-v1/GEN-SIM-RECO) were compared with CMSSW_2_2_X tt+jets samples (/TauolaTTbar/Summer08_IDEAL_V11_redigi_v2/AODSIM) at PAT level :
- CMSSW_3_1_2 tt+jets samples available (/TTbar/Summer09-MC_31X_V3_preproduction_312-v1/GEN-SIM-RECO)
- statistics : 100k for Summer08 / 70k for Summer09 [311]
- no selection cuts
- same JEC factors (Summer08Redigi)
- jet algorithm : SISCone05
- GlobalTag : IDEAL_V11::All/STARTUP31X_V1::All [mistake done only for 311 but do not change results]
- Results 311 vs 22X: http://beaucero.home.cern.ch/beaucero/MCValidation_ReducedPages.ps
- Results 312 vs 22X: http://beaucero.home.cern.ch/beaucero/MCValidation22Xvs312_ReducedPages.ps
- Results 312 vs 311: http://beaucero.home.cern.ch/beaucero/MCValidation311vs312_ReducedPages.ps
- cfg file to be used :
==== Comments for 22X vs 311 comparison: ====
- Plots 1-34 made at generator level.
- differences compatible within statistical uncertainties
- Plots 35-40 made at PAT level with all jets :
- plot 36 : differences in jet towers area for low values
- plot 37 : differences in fraction of jets with electromagnetic energy fraction less than 0.2
- plot 38 : differences in fraction of jets with hadronic energy fraction equal to 1
- plots 39/40 : differences in low values for the maximum energy deposited in ECAL/HCAL towers
- Plots 41-46 made at PAT level with jets matching primary quarks
- differences observed in previous plots less pronounced but there are still some discrepancies :
- plot 43 : differences in fraction of jets with electromagnetic energy fraction equal to 0
- plot 44 : differences in fraction of jets with hadronic energy fraction equal to 1 (correlated to diff in plot 43)
- Plots 47-54 made at PAT level with all jets
- plot 48 : differences in eta position of the jet calotowers
- Plots 55 and 56 made at PAT level with the "trackCountingHighPurBJetTags" b-tag algorithm
- plot 55 made with all the jets : no difference
- plot 55 made with jets matching b-quarks : no difference
- Plots 57-86 made at PAT level
- plots 57-60 : differences in MET and MET phi values around zero (forward hadronic calorimeter)
- plots 61-75 : MET kinematics with different un-correction factors (see http://cms-service-sdtweb.web.cern.ch/cms-service-sdtweb/doxygen/CMSSW_3_1_0/doc/html/db/deb/classpat_1_1MET.html#a00128d224f81232e649ae61a1117352)
- plot 63 : differences for (ALL) uncorrected ET sum (especially between 20 and 50 GeV)
- plot 64 : (ALL) uncorrected MET phi not flat in CMSSW_3_1_1
- plot 68 : differences for (JES) uncorrected ET sum (especially around 30 GeV)
- plot 69 : (JES) uncorrected MET phi not flat in CMSSW_3_1_1
- plot 74 : (MUON) uncorrected MET phi not flat in CMSSW_3_1_1
- plots 76-84 : event energy fraction in the various sub-detectors (http://cms-service-sdtweb.web.cern.ch/cms-service-sdtweb/doxygen/CMSSW_3_1_0/doc/html/db/deb/classpat_1_1MET.html#8a4db29a4ae9fd639c3e87b4621a6930)
- plot 76 : event electromagnetic energy fraction shifted to lower values in CMSSW_3_1_1
- plot 78 : event electromagnetic energy fraction shifted to lower values in CMSSW_3_1_1 (ECAL end-cap)
- plot 80 : correlated with plot 76
- plot 83 : event hadronic energy fraction shifted to higher values in CMSSW_3_1_1 (forward HCAL)
- Plots 87-91 made at generator level
- plot 87-91 : no significant difference for the generated MET
- Plots 87-91 made at generator level
- Plots 92-106 made at PAT level with global muons (not matching the muon from top->W decay)
- InnerTrack 92-94 :
- plot 92 : less valid hits in silicon fit for CMSSW_3_1_1 global muons
- OuterTrack 95-96 :
- plot 95 : buggy plot (the one for the lost hits is not in the ps file)
- GlobalTrack 97-99 :
- plot 99 : normalized global fit Chi² shifted to lower values for CMSSW_3_1_1 global muons
- Other quantities 100-106 :
- plot 100 : differences in calo-compatibilty
- plot 101 : differences in TrackIso variable for low values
- Plots 107-120 made at PAT level with global muons (matching the muon from top->W decay)
- GlobalTrack 111-113 :
- plot 113 : same observation as for plot 99
- Other quantities 114-120 :
- plot 114 : differences in calo-compatibilty
- plot 115 : no difference for TrackIso variable between top matched global muons anymore (see comments on plot 101)
- Plots 92-106 made at PAT level with global muons (not matching the muon from top->W decay)
- Plots 121-133 made at PAT level with all vertices
- plots 128/129 : more tracks associated to the primary/all vertices in CMSSW_3_1_1
- Good agreement between two releases for all plots
- Plots 134-186 made at PAT level
- plots 134-140 : Jet kinematic variables (all jets)
- plot 134 : higher jet multiplicity for CMSSW_3_1_1
- plot 137 : different eta distribution
- plot 138 : different energy distribution for low values
- plot 139 : different theta distribution
- plots 141-146 : Muon kinematic variables (all muons)
- plot 144 : different eta distribution for |eta|>2.0
- plots 141-146 : Muon kinematic variables (all muons)
- plots 147-153 : MET kinematic variables (all mets)
- plot 148 : MET spectrum shifted to higher values in CMSSW_3_1_1
- plot 153 : MET phi spectrum not flat in CMSSW_3_1_1
- all other plots : compatible within statistical uncertainties
- plots 147-153 : MET kinematic variables (all mets)
- plots 154-160 : Jet kinematic variables (jets matching primary quarks)
- all plots : compatible within statistical uncertainties
- plots 154-160 : Jet kinematic variables (jets matching primary quarks)
- plots 161-173 : Muon kinematic variables (muons matching muon from top sub-decay)
- all plots : compatible within statistical uncertainties
- plots 174-186 : MET kinematic variables (met matching primary neutrino from top sub-decay)
- all plots : compatible within statistical uncertainties
- plots 161-173 : Muon kinematic variables (muons matching muon from top sub-decay)
- Plots 187-198 : require at least 4 jets with pT > 30 and |eta|<2.5, require at least 1 muon with |eta|<2.5 which is isolated according to the veto cone isolation requirement (vetoEM<4, vetoHad <6 GeV) - basically a rough attempt to reconstruct the semi-leptonic ttbar event
- plots 189-197: all 4 jets should be matched to partons:
- plots 187-189: (quarkEnergy-jetEnergy)/jetEnergy
- plots 190-191: rank and pT of the lowest pT jet that is matched to a parton
- plots 192-197: generated and reconstructed leptonic and hadronic W and top masses
- plot 198 when the 4 quarks are not all 4 matched to the partons: check how many of them are not matched (multiplicity)
- all plots : compatible within statistical uncertainties
- plots 189-197: all 4 jets should be matched to partons:
- Plots 187-198 : require at least 4 jets with pT > 30 and |eta|<2.5, require at least 1 muon with |eta|<2.5 which is isolated according to the veto cone isolation requirement (vetoEM<4, vetoHad <6 GeV) - basically a rough attempt to reconstruct the semi-leptonic ttbar event
==== Additional Comments for 22X vs 312 comparison [and comments are also valid for 311 vs 312 comparison] ==== Note that previous comments are still valid except when quoted
- plot 37 : differences in fraction of jets with electromagnetic energy fraction less than 0.2 --> Difference is reduced in 312
- plot 38 : differences in fraction of jets with hadronic energy fraction equal to 1 --> Difference is reduced in 312
- plot 43 : differences in fraction of jets with electromagnetic energy fraction equal to 0 --> No More Discrepancies in 312
- plot 44 : differences in fraction of jets with hadronic energy fraction equal to 1 (correlated to diff in plot 43) --> No More Discrepancies in 312
- plot 63 : differences for (ALL) uncorrected ET sum (especially between 20 and 50 GeV) --> Difference is reduced in 312
- plot 64 : (ALL) uncorrected MET phi not flat in CMSSW_3_1_1 --> More flat in 312
- plot 68 : differences for (JES) uncorrected ET sum (especially around 30 GeV) --> Difference is reduced in 312
- plot 69 : (JES) uncorrected MET phi not flat in CMSSW_3_1_1 --> More flat in 312
- plot 74 : (MUON) uncorrected MET phi not flat in CMSSW_3_1_1 --> More flat in 312
- plots 76-84 : event energy fraction in the various sub-detectors
- plot 76 : event electromagnetic energy fraction shifted to lower values in CMSSW_3_1_1 --> Shifted back in 312
- plot 78 : event electromagnetic energy fraction shifted to lower values in CMSSW_3_1_1 (ECAL end-cap) --> Shifted back in 312
- plot 80 : correlated with plot 76 --> Shifted back in 312
- plot 148 : MET spectrum shifted to higher values in CMSSW_3_1_1 --> Shifted back in 312
- plot 153 : MET phi spectrum not flat in CMSSW_3_1_1 --> More flat in 312
== Conclusion ==
312 is better than 311, 312 can be considered has been validated by Brussels groups as the remaining discrepancies have been explained by change of code in tracker etc.