-- AmnonHarel - 17 Feb 2009


See index in AmnonHarelDataMixing

Testing Mike's DataMixer, circa 1-19 Feb. Goal was to see a plot of HPD discharge noise overlayed near / over MC signal. The closest we could get with "version 3" of the test file (due to reasons described below) is: sav3mix24.gif

To understand this event, look at the attached pure MC, overlayed data, and the HPD noise analysis of the data and the mixed MC. The HPD analyzed is the one chosen by my HPD discharge noise ID, which uses only the geometry of the HCal energy in eta-phi space.

Conclusions (on the overlaying process, and reconstruction)

  • made two new reconstruction configs recoMixed_cff.py and RecoLocalTrackerNoZeroSuppression_cff.py (exact locations below) to be able to reconstruct overlayed data
  • this CRAFT data should not be be mixed - onward to V4 in AmnonHarelDataMixingRound4


  • work area:
    • Old one is /uscms/home/aharel/CMSSW_2_1_10/src/SimGeneral/DataMixingModule
    • New one described below /uscms/home/aharel/CMSSW_2_2_3/src/
  • Mike's notes on reco-ing:
    • CMSSW 2_2_3 was used for making the overlays, but caused problems when reco-ing MC.
    • He then generated some new MC for 2_2_3, and that's what is in the _v3 files .
    • Mike's reco config is /uscms/home/mikeh/work/CMSSW_2_2_3/src/SimGeneral/DataMixingModule/python/datamixertest_RAW2DIGI_RECO_STARTUP.py
      • But that one doesn't use the mixer output !?
      • this one, in the same location, does DigiRecoExampleMC_STARTUP_cfg.py, and it looks to do so exactly like my config from round1, so I'm using that /uscms/home/aharel/test/recoMixed_cff.py.
    • In 2_1_8/9, Mike "had to turn off Preshower5x5 Clustering in the EndCap". This is not an issue for my test.
    • Mike also run into the problem of the siStripZeroSuppression, since the mixer (currently) outputs SI digis that replace the zero suppressed digis, instead of the lower level virginraw digis. I copied and renamed his config to /uscms/home/aharel/CMSSW_2_2_3/src/SimGeneral/DataMixingModule/python/RecoLocalTrackerNoZeroSuppression_cff.py
  • data area: contains both .log and .root files /uscms_data/d1/aharel
  • v4 MC events: /uscms_data/d1/mikeh/QCD_Pt_50_80_cfi_GEN_SIM_DIGI_L1_DIGI2RAW_NoNoise.root
    • examined using gdQCD4.py --> gdQCD4.log * this file contains neither reco. calo jets, nor calorimeter digis
  • v3 MC events: /uscms_data/d1/mikeh/QCD_Pt_50_80_cfi_GEN_SIM_DIGI_L1_DIGI2RAW.root
    • examined using gdQCD3.py --> gdQCD3.log
      • this file contains neither reco. calo jets, nor calorimeter digis
    • reco-ed & translated into my format using prepQCD3.py --> prepQCD3.log, prepQCD3events.root (in CMSSW format) & prepQCD3.root (in my format)
  • early MC events: /store/relval/CMSSW_2_1_8/RelValQCD_Pt_80_120/GEN-SIM-DIGI-RAW-HLTDEBUG-RECO/STARTUP_V7_v1/0003/0A0241FB-9182-DD11-98E1-001617E30D40.root
    • used copyQCD2.py to copy 500 events into QCD2events.root
      • odd, only 450 events there.
    • examined using gdQCD2.py --> gdQCD2.log
      • file contains reconstructed jets and simCalDigis (i.e. simEcalDigis and simHcalDigis)
    • translated to my sc.c format using prepQCD2.py --> prepQCD2.log
  • Non-collision data events: /store/data/Commissioning08/Cosmics/RAW/v1/000/070/664/20A57DB9-79AF-DD11-95AB-000423D99BF2.root Mike also provided "Direct Reco file":
    • /uscms_data/d1/mikeh/promptrecoCosmicsDigis_2_2_3.root
      • examined using gdData2_2_3.py --> gdData2_2_3.log
      • this file contains reconstructed jets, but no calorimeter digis
      • translated to my sc.c format using prepData2.py --> prepData2.log
      • no clear HCal anomolies in the events I translated (first 100). Most suspicious (no EM energy nearby) are #79, #233, #343 (physical?), #454, #455, #460. The corresponding CMSSW event #s are 21017408, 21029876 (--> MC#24), 21038460, 21045212, 21045448, 21045688.
      • physical looking energy:
        • EM energy nearby: event #214, #385, #405, i.e., 21028404, 21040920 (--> MC#176), 21042488 (--> MC#196).
        • muon shower? in event #216, i.e., 21028474 (-->MC#7).
      • all the interesting (hardest in event and noise-like) energy is in the first 3 digis - what trigger and/or noise type is this?
      • Dima 's warning: this looks like a high rate run (verified in run registery) with only a few time slices read out, and with aggressive 0-suppresion.
    • /uscms_data/d1/mikeh/promptrecoCosmicsDigis_2_1_8.root
      • examined using gdData2_1_8.py --> gdData2_1_8.log
      • this file contains calorimeter digis, but no reconstructed jets
  • Overlayed events: 500 such events were provided by Mike in two formats
    • "Original Mixed Digis": /uscms_data/d1/mikeh/CRAFT_on_QCD_Digis_v3.root
      • examined using gdOverlayDigi.py --> gdOverlayDigi.log
      • this file contains mixed calorimeter digis (with the usual mixer names = reco names), but no reconstructed jets
    • "Reco'ed overlay": /uscms_data/d1/mikeh/CRAFT_on_QCD_Reco_SV7_v3.root
      • examined using gdOverlayReco.py --> gdOverlayReco.log
      • this file contains reconstructed jets, but no calorimeter digis
    • Also available there is CRAFT_on_QCD_eventlist.txt

  • new work area:
    • was set up with:
cmsrel CMSSW_2_2_3
cd CMSSW_2_2_3/src
addpkg RecoJets
cd RecoJets/JetAnalyzers/interface/
ln -s ~/test/AHtest.h AHtest.h
cd ../src
ln -s ~/test/AHtest.cc AHtest.cc
cd ..
mkdir python
ln -s ~/test/AHtest_cfi.py python/AHtest_cfi.py
mv BuildFile BuildFile_org
ln -s ~/test/BuildFile BuildFile
scram b

  • ...
  • HCal energies - data from V3 - event #233:
Topic attachments
I Attachment History Action Size Date Who Comment
GIFgif sav3data233.gif r1 manage 17.5 K 2009-02-20 - 15:29 AmnonHarel HCal energies - data from V3 - event #233
PNGpng sxv3data233.png r1 manage 31.7 K 2009-02-20 - 15:30 AmnonHarel HCal noise analysis - data from V3 - event 233
Edit | Attach | Watch | Print version | History: r10 < r9 < r8 < r7 < r6 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r10 - 2009-08-12 - AmnonHarel
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback