160 likes | 272 Views
Jet Validation of Summer11 MC. Artur Apresyan Caltech Joanna Weng ETH Zurich Kittikul Kovitanggoon Texas Tech University. Summer11 wrt Spring11_ S2. The datasets are Summer11 MC processing with CMSSW_4_2_2 /TT_TuneZ2_7TeV-pythia6-tauola/Summer11- PU_S3 _START42_V11-v2/GEN-SIM-RECO
E N D
Jet ValidationofSummer11 MC ArturApresyan Caltech Joanna Weng ETH Zurich KittikulKovitanggoon Texas Tech University
Summer11 wrt Spring11_S2 • The datasets are • Summer11 MC processing with CMSSW_4_2_2 • /TT_TuneZ2_7TeV-pythia6-tauola/Summer11-PU_S3_START42_V11-v2/GEN-SIM-RECO • Spring11 MC processing with CMSSW_4_1_4 • /TT_TuneZ2_7TeV-pythia6-tauola/Spring11-PU_S2_START311_V2-v2/GEN-SIM-RECODEBUG • Both datasets have inTime and OOT PU, but the Spring11 sample mistakenly was produced with 25 ns bunch-spacing • In both samples, the inTime and OOT PU have the “Flat10” distribution • https://indico.cern.ch/getFile.py/access?contribId=3&resId=1&materialId=slides&confId=136336 • Summer11 has 50 ns bunch-spacing • Details about various PU configurations are at: • https://twiki.cern.ch/twiki/bin/viewauth/CMS/PileupInformation
Generator Level Jets Summer11 Spring11 • The jet parameters look identical on both Summer11 and Spring11_S2 • Generator level includes only particles from primary collision, no pile-up
Reco jet Multiplicity • This and the following slides will show comparison for AK5 CaloJets • Conclusions for other types of jets are the same • Higher jet occupancy in Spring11_S2, which is expected from OOT PU that is more pronounced in Spring11 (25ns pileup) • Number of jets is higher in Spring11_S2 sample, but they are mostly soft jets from different OOT PU Summer11 Spring11 # of jets, |η|≤1.3, pT>10 GeV Jets Pt # of jets, |η|>1.3, pT>10 GeV
Jet pT scale (pTcalo/pTgen) Summer11 Spring11 PT scale, |η|≤1.3 PT scale, 1.3<|η|<3.0 PT scale, 3<|η|<5.0 • Jet scale in Spring11_S2 is higher than that in Summer11 in all detector regions • Most pronounced in jets up to ~100 GeV • The difference is bigger in Endcaps, and most pronounced in the HF region • Again, expected from 25 ns OOT PU in Spring11_S2
Jet Eta and Phi distributions Summer11 Spring11 Jets η, PT>10 GeV Jets φ, PT>10 GeV • Flat in Phi, the “horns” in eta are more pronounced in Spring11_S2 likely due to 25ns
Summer11 wrt Spring11_S1 • The datasets are • Summer11 MC processing with CMSSW_4_2_2 • /TT_TuneZ2_7TeV-pythia6-tauola/Summer11-PU_S3_START42_V11-v2/GEN-SIM-RECO • Spring11 MC processing with CMSSW_4_1_4 • /TT_TuneZ2_7TeV-pythia6-tauola/Spring11-PU_S1_START311_V1G1-v1/AODSIM • Here we are comparing to Spring11_S1 sample which has NOOOT PU wrt Summer11 sample. • Summer11 sample is the same as shown on previous pages, with 50ns OOT PU • Detailed comparisons at: • http://highenergy.phys.ttu.edu/~keng/validation/plots/TT_Summer11_wrt_Spring11_S1_area/
Reco jet Multiplicity Summer11 Spring11 • This and the following slides will show comparison for AK5 CaloJets • Conclusions for other types of jets are the same • Slightly higher jet occupancy in Summer11 samples, effect of OOT PU Jets Pt # of jets, |η|≤1.3, pT>10 GeV # of jets, |η|>1.3, pT>10 GeV
Jet Eta and Phi distributions Summer11 Spring11 Jets η, PT>10 GeV Jets φ, PT>10 GeV
Jet pT scale (pTcalo/pTgen) Summer11 Spring11 PT scale, |η|≤1.3 PT scale, 1.3<|η|<3.0 PT scale, 3<|η|<5.0 • Jet scale in Summer11 is higher than that in Spring11_S1, in all detector regions • Expected for barrel and endcap, but not immediately clear why different in HF • HF energy reconstruction is done with 2TS, should be insensitive to 50ns OOT PU • Maybe PileUp (inTime) configuration is different between S3 and S1?
Energy in subdetectors Summer11 Spring11 Unexpected difference in HF Energy in EB Energy in EE Energy in HF Energy in HB Energy in HE
Jet pT response (pTcorrected/pTgen) Summer11 Spring11 PT scale, |η|≤1.3 PT scale, 1.3<|η|<3.0 PT scale, 3<|η|<5.0 • After applying the L2L3 jet energy corrections the differences are not as large between two samples • But it is still not clear why there is difference in previous page…
4_3_0_pre4 • The energy change for the JPT algorithm. 4_3_0_pre4 4_3_0_pre2 PT scale, |η|≤1.3 PT ratio • This discrepancies are understood to be caused by a change in JPT algorithm. • Starting from some 4_3_0_preX we use UseZSP = cms.bool(False) in agreement with the new JEC structure for JPT. • ZSP corrections should go into L1JPTOffset. Staring in 4_3_0_pre6 release.
Conclusions • The Jet validation for normal release is moving smoothly • Spring11_S2 MC has more jets than Summer11 MC, but they are mostly soft jets from different OOT PU (25 ns OOT PU in Spring11_S2 and 50 ns OOT PU in Summer11) • Jet pT response is higher in Spring11_S2 than Summer11. This again is expected from different OOT PU • Generator level jets are identical in both datasets • Spring11_S1 sample looks OK in general, but some features need further investigatio • Response is OK, Spring11_S1 can likely be used for analysis, we are following this up with HCal group