Purpose of this twiki

Trigger objects monitoring

Quickstart Guide

See the plots in the Gui

Collision Reference Plots

Last Update: June 23 2010

Current Collision Reference: 137028

NOTE: Please choose to look at these plots in the /StreamExpress/ dataset. If you choose another dataset, the plots will look different due to trigger bias.

01 HLT_Mu3 Pt

HLT_Mu3 Pt
Description: Shows Eff vs Pt. Evaluation: Check for significant deviations against reference

02 HLT_Mu3 Phi vs Eta

HLT_Mu3 Phi vs eata
Description: Shows Eff vs Pt. Evaluation: Check for significant deviations against reference

03 HLT_Mu5 Pt

HLT_Mu5 Pt
Description: Shows Eff vs Pt. Evaluation: Check for significant deviations against reference

04 HLT_Mu5 Phi vs Eta

HLT_Mu5 Phi vs Eta
Description: Shows Eff vs Pt. Evaluation: Check for significant deviations against reference

05 HLT_Mu9 Pt

HLT_Mu9 Pt
Description: Shows Eff vs Pt. Evaluation: Check for significant deviations against reference

06 HLT_Mu9 Phi vs Eta

HLT_Mu9 Phi vs Eta
Description: Shows Eff vs Pt. Evaluation: Check for significant deviations against reference

See Some RelVal Plots

Plots for a few RelVal samples in 3_1_X releases are available from the following link: http://slaunwhj.web.cern.ch/slaunwhj/MuonOfflineDQMPlots/.

Software structure

The software is structured according to the following diagram:

The inputs are a list of triggers to monitor and a list of selections. The selections apply to both muon quantities (z0, d0, pt, etc) and event quantities (require a certain trigger to fire). The selections determine

• Which events to compare
• Which HLT and RECO muons to compare

Which plots are produced?

The muon offline DQM produces ~1000 histograms using an average of 28 bins per histogram. The breakdown of the plots is:

# Last updated 2009-09-18

8 Triggers (HLT_L1Mu, HLT_L1MuOpen, 
                  HLT_L1Mu20, HLT_L2Mu9, 
                  HLT_L2Mu11, HLT_IsoMu3, 
                  HLT_Mu3, HLT_Mu5, HLT_Mu9, 
                  HLT_DoubleMu3, HLT_DoubleMu0, 
                  HLTL1DoubleMuOpen) 

3 Selections (barrel, endcap, all)
~50 plots per selection

For each selection, the plots are 

24 Efficiency related plots: 
     D0Beam, Z0, Charge, eta, phi, pt, pt_max, pt_exactly_one_muon,
     8 distributions * 3 histos/distrib = 24 plots
9   Fake related plots: pt, eta, phi (ONLY produced for HLT Triggers)
1 Charge Flip
2 isolation
4 track wrt (0,0,0) Z0Beam, d0 detector 
4 Beamspot track parameters: D0, Z0 eff from beamspot
2 Phi Vs Eta
3 Resolution: Pt, eta phi

To see the most current selections, look at the following configuration files: Collisions Cosmics

Where are the plots located in the DQM plot directory structure?

The base directory is DQMData/HLT/Muon/Distributions/. There is a separate directory for each trigger. Each trigger directory has a subdirectory for each muon selection. So the path for trigger "HLT_Mu9", considering selection "allMuons" is DQMData/HLT/Muon/Distributions/HLT_Mu9/allMuons.

Run The Code

Follow these steps to get started running the DQM right away.

1. Log into one of the pool machines

          # for releases >= 34X, you need to have SL5 
          ssh lxplus5
          

2. Check out your favorite release of CMSSW that is no older than 3_1_0_pre4. We will refer to your release as CMSSW_X_Y_Z.

         Example:
         scram project CMSSW CMSSW_3_5_0_pre3
         

3. Check out the DQMOffline Package using cvs

   cvs co -r CMSSW_X_Y_Z DQMOffline/Trigger

            Example:
             cvs co -r CMSSW_3_5_0_pre3 DQMOffline/Trigger

4. Build your version of CMSSW with scram b

            scram b -j 4
          

5. Run the DQM analysis step.

   Example:
   cd DQMOffline/Trigger/test/
   
   # You can change the input file to be whatever you like
   
   cmsRun muonTest_cfg.py print \
   files='/store/relval/CMSSW_3_4_0_pre1/RelValTTbar/GEN-SIM-RECO/MC_31X_V9-v1/0008/2C8CD8FE-B6B5-DE11-ACB8-001D09F2905B.root' \
   output='file:/afs/cern.ch/user/s/slaunwhj/scratch0/EDM_ttbar_340pre1.root' \
   maxEvents=100 > & ! ./test_340pre4.log &
   
   

6. Run the DQM harveting step

       Example:
       # make sure that the input files in the script match the output from your previous script
        cmsRun muonPostProcessor_cfg.py print files='file:/afs/cern.ch/user/s/slaunwhj/scratch0/EDM_ttbar_340pre1_numEvent100.root' \
        outputDir='/afs/cern.ch/user/s/slaunwhj/scratch0/' \
        workflow='/DQMGeneric/PostProcess/Testv1' maxEvents=-1 > & ! post.log
      

7. Look at the histograms in the TBrowser. The histograms are in DQMData/HLT/Muon/Distributions/. Each trigger has it's own sub-directory.

Four vector dqm source

There exists a generic monitoring tool for HLT performance giving insights to the performance of the muon triggers. Documentation is available under SWGuideFourVector . The directory for the plots is HLT/FourVector/client/ trigger path /custom-eff

Local reconstruction monitoring

There is currently no specific local reconstruction monitoring in the offline DQM.

PAG trigger monitoring

Top PAG Muon Trigger Monitoring

The top group is concerned about monitoring muon trigger efficiencies in high-occupancy events. The have supplemental monitoring under development documented here: SWGuideMuonHLTTopPAGDQM.

Tag and Probe Muon Trigger Monitoring

The B-physics group is interested in tag-and-probe with low-mass di-muon resonances such as J/Psi and Upsilon. They have supplemental monitoring under development documented here: SWGuideMuonHLTBPAGDQM.

Muon HLT Offline Analysis

There are some instructions for offline analysis that extends beyond offline DQM available on the SWGuideMuonHLTOfflineAnalysis.

-- JasonSlaunwhite - 29 Apr 2009