GOAL

• calibrate and optimize CSC Track Finder momentum measurement for 2015 run

TASKS

Establishing readiness of necessary tools:

• Calibration repeat should be done on Monte Carlo only, unless it is found that the calibration currently being used is derived from data.
• If single-momentum muon gun Monte Carlo samples do not exist, generate and produce these samples for the calibration study
• Repeat study done by Anna using the official L1TriggerDPG code
• Establish ability to measure 1/pT resolution from residuals and turn-on curve shape
• Establish ability to measure plateau efficiency from turn-on curve shape
• Establish ability to calculate rate reduction factor on min Bias data

• perform the above calibration analysis using the new magnetic field map and un-ganged ME1/1a chambers (involves re-simulating MC samples)
• report efficiencies, resolution and rate reduction factor, expected performance relative to that of current system

• study distribution of track modes, identify those which contribute the most rate
• for track modes which contribute the most to the trigger rate, establish most important variables to be used in pT assignment
• investigate if path length correction (front vs rear chambers) can be used to improve pT assignment
• optimize use of address space to maximize information from important variables, derive improved look-up tables
• report efficiency, resolution and rate reduction for the improved tables

• investigate if GMT muon combination algorithm is limiting the power of the muon trigger by giving too much weight to RPC tracks
• optimize GMT pT assignment algorithm to maximally reduce rate while preserving efficiency

• present results, as available, in the daily trigger meetings, and in the Weekly CMS-wide CSCTF meetings
• as appropriate and approved in CSCTF meetings, report findings in L1 Trigger meeting
• document study as a Technical Note

Immediate Technical To Do List

• getting to know the CMS L1 trigger system and CSCTF jargon, see recommended Documents
• getting to know ROOT tool at root.cern.ch
• play with the ntuples created by Carlo Battilana

Code

• L1TriggerDPG Repository: http://cmssw.cvs.cern.ch/cgi-bin/cmssw.cgi/UserCode/L1TriggerDPG/
• Quick howto at https://twiki.cern.ch/twiki/bin/viewauth/CMS/CSCTFPromptAnalysis

# Setup base environment
mkdir code
cd code

# Check-out CMSSW code and compile it
export SCRAM_ARCH=slc5_amd64_gcc462
cmsrel CMSSW_5_3_8_patch3
cd CMSSW_5_3_8_patch3/src/
cvs co UserCode/L1TriggerDPG
cmsenv
scramv1 b -j 9

# Get the ntuples and the test code
cd UserCode/L1TriggerDPG/macros/CSCTF/
#copying may take a while
cp /afs/cern.ch/work/d/digiovan/public/forDavidR/* .
chmod 755 setup

# run the example
source setup.sh
root -l runTestForDavid.C++

# this example will create a root file and save into it the distribution of single muon pT for global muons 

Suggestions:

   modify the code to draw also the muon eta, phi and repeat (pt, eta, phi) also for standalone and tracker muon.
   plot the number of recontructed muon (with pt>5 GeV) per event 
   scatter plot of global vs standalone muon pt
   evaluate the pt resolution of standalone vs global muon ( rms of the following distribution (gm_pt - stm_pt) / gm_pt, determine also the mean and sigma of the gaussian fit to the previous distribution and understand the results)
   do the same guassian fit for muon eta and compare to the pt case, make sure you undertand the difference, if any    

* Select only global muon in the CSCTF region (apply the correct selection in |eta|)
* Given a global muon, loop over the CSCTF tracks and plots the eta and phi difference between CSCTF tracks and global reco muons
* For all the CSCTF tracks please plot the mode, quality, pt and number of stubs
* For all the CSCTF tracks please plot the mode vs quality, mode vs number of stubs

* Compute the delta phi, delta eta, delta R difference using rchPhi, rchEta for the muons instead of trEta or trPhi
* Compare on the delta phi, delta eta and delta R with the previous ones: put the 2 delta phi distributions in a single canvas. Same for delta eta and delta R. Which variables give you the narrower distribution?
a. Prepare a 1-D histograms with the pT of the global muon with dynamic binning for only muon candidates which are in the CSCTF region and have DR<0.2 (use the DR which is narrower one)
b. Prepare a 1-D histograms with the pT of the global muon with dynamic binning for only muon candidates which are in the CSCTF region and have DR<0.2 and Quality is >= 2
c. Prepare a 1-D histograms with the pT of the global muon with dynamic binning for only muon candidates which are in the CSCTF region and have DR<0.2 and Quality is >= 2 and the CSCTF  pT > 5.
d. Prepare a 1-D histograms with the pT of the global muon with dynamic binning for only muon candidates which are in the CSCTF region and have DR<0.2 and Quality is >= 2 and the CSCTF pT > 7.
e. Prepare a 1-D histograms with the pT of the global muon with dynamic binning for only muon candidates which are in the CSCTF region and have DR<0.2 and Quality is >= 2 and the CSCTF pT > 10.
* Try to construct a TGraphAsymmErrors obtained by the ratio of 2 histograms: numerator is the pT plot with DR<0.2 and Quality >=2  and CSCTF pT > 7. (plot d.) while the denominator is the  pT of the global muon with DR<0.2 and Quality >=2 (plot b.)
* Fit the efficiency vs pt with an ErrF function and comment the results
* apply a reasonable cut on global muon pt (> csctf pt) and study the efficiency vs phi. Please comment on the result
* apply a reasonable cut on global muon pt (> csctf pt) and study the efficiency vs eta. Please comment on the result
* Let's go back to the the efficiency vs pt plot. Are we sure the bins we are choosing are really the optimal boundaries? Can you show the global vs CSCTF pt muon resolution in each bin? Try different bin boundaries until the resolution is really the same for all.

* Test the file in /afs/cern.ch/work/d/digiovan/public/forDavidR/L1Tree_Run2012C_SingleMuOpen_RAW-RECO_GP.root. Does it have more entries than the original one? (Do not sum the two as in principle this file contains the data from the original one)
* Try to fit the turn on curves with a function like TF1* fitPTfunc = new TF1("fitPTfunc", "[0]*TMath::Erf([1]*x-[2])+[3]", 2., 140.)
* Try to fit the phi distribution with a straight line
* Get to know how to use crab commands, using the tutorial https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookCRABTutorial. You may also try to run and copy some files on run2012C on alachua using my scripts
* Can you study also the resolution (tr_pt-csctf_pt) for the 3 quality values and in 4 ranges of eta: 0.9=<|eta|<=1.2, 1.2<|eta|<=1.6, 1.6<|eta|<=2.1, |eta|>2.1. You can also mix them up in case you have enough stats, e.g. Q>=2 and 1.6<|eta|<=2.1. The idea is to see which region or quality cut provides the best resolution (i.e. how far/close is the transverse momentum prediction of the CSCTF w.r.t. the best momentum measurement CMS can perform)

Submit Using Crab

• Request your permission to submit jobs on the grid at https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookRunningGrid (follow "Get a Grid certificate and the registration to CMS VO")
• Tutorial: https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookCRABTutorial
• Since David cannot save the ntuples on CAF (need at least 6 months at CERN), he will be able to save them at UF and then copy on lxplus

# Setup base environment
mkdir code
cd code

# Check-out CMSSW code and compile it
export SCRAM_ARCH=slc5_amd64_gcc462
cmsrel CMSSW_5_3_8_patch3
cd CMSSW_5_3_8_patch3/src/
cvs co UserCode/L1TriggerDPG
cmsenv
scramv1 b -j 9

# create your environment for crab
cd UserCode/L1TriggerDPG/test
mkdir crab
cd crab

# this is an example to run over data on the CIEMAT Tier-2
# from Commissioning Run2012C PromptReco-v2
cp -r /afs/cern.ch/work/d/digiovan/public/forDavidR/crabExampleFiles/* .

# If you look into the crab.cfg you will see it asks the jobs to be run on CIEMAT
# on the dataset Commissioning Run2012C PromptReco-v2 and store in digiovan's area
# /cms/data/store/user/digiovan/csctfStudies/ in alachua.ihepa.ufl.edu
# you will need to change location to store the files in a different location and probably the dataset to run over

Some info about the previous work from Anna

• Presentation available at https://twiki.cern.ch/twiki/pub/Main/CsctfPtLutsSummerStudy/kropiv_2011.05.30approval.pdf

Documents

CMS Trigger System:

• trigger TDR: http://cmsdoc.cern.ch/cms/TDR/TRIGGER-public/trigger.html (in particular chapter 1 and 12)
• muon TDR: http://cds.cern.ch/record/343814/
• CSC-CSCTF conventions: http://cms.cern.ch/iCMS/jsp/openfile.jsp?type=IN&year=2007&files=IN2007_024.pdf

• https://twiki.cern.ch/twiki/bin/view/CMSPublic/WorkBookBasicLinux
• https://twiki.cern.ch/twiki/bin/view/CMS/WorkBook
• https://twiki.cern.ch/twiki/bin/view/CMS/Tutorials
• http://root.cern.ch/drupal/

• Definition of Tight muon, https://twiki.cern.ch/twiki/bin/view/CMSPublic/SWGuideMuonId#Tight_Muon

Useful Info

• Increase AFS quota, http://account.web.cern.ch/account/Services/MyServices.aspx?service=Linux%20and%20AFS

• Slides_For_Project.pptx