Running on the CMSSW_1_6_12 fast sim soups for Top analysis

The following contains some notes on how to run CMSSW_1_6_12 for the semileptonic (muon) ttbar channel.

Dataset

• The completion of the CMSSW_1_6_12 fast sim production was announced here on 2008-06-28.
• tt is in Chowder soup (see also the Soup definitions).
• See this link for the chowder files at CERN and here for the DBS entry.
  • note that CERN only has 624866 events (according to DBS; my job ran over 628658 events), other sites have the full 11896860 events.
  • running on this sample, I see 16261 semileptonic (e: 5361, mu: 5495, tau: 5405), 16266 fully hadronic and 4099 fully leptonic events (36626 events in total, 6.3 GB including TQAF generator information and PF jets with two clustering algorithms = 179 kB / event)

Selection of tt events from the soup

Creation of TQAF objects

TtGenEvent

TtGenEvent

TtGenEventReco

Note that this seems to require the new collection genParticles while in CSA07 samples the old genParticleCandidates collection is produced. See also this link.

• To print the list of generated particles, see here. This seems also to work with e.g.

  InputTag src = decaySubset

  to print the list of decay particles selected for creating the TtGenEvent.

Finding reconstructed muons

Finding reconstructed jets

Particle flow

Particle candidates

• PFCandidate in LXR

Jets (from particle flow candidates ?)

• the object (in the standard files) is particleFlowJetCandidates (or recoCandidatesOwned_particleFlowJetCandidates__PROD.obj with it's non-aliased name), but it is not clear to me what is in there...
• there seem to be configuration files to do jet reconstruction: RecoParticleFlow/PFProducer/data/particleFlow.cff, RecoParticleFlow/PFProducer/data/particleFlowJetCandidates.cfi
  • It looks like it's basically a matter of setting a different type of input particles to the jet algorithm

    InputTag src = particleFlowJetCandidates

Example plots

Note that TtGenEvent::lepton() only returns leptons, but not anti-leptons. However, it looks like most of the quantities of the returned lepton are set to zero if there is no lepton (as opposed to an antilepton) in the W decay, so probably one can safely add the quantity returned by lepton().xxx and leptonBar().xxx when plotting.

Physics

• Physics analysis summaries for top: here
  • PAS for semileptonic (muon) xsect with 10/pb

Number of expected events in PTDR analysis

PTDR

Table 8.4 from PTDR vol II:

Python

• setup python in your CMSSW area (as described here

cvs co PhysicsTools/PythonAnalysis/python
export PYTHONPATH=$CMSSW_BASE/src/PhysicsTools/PythonAnalysis/python:$PYTHONPATH

Configuring from python

• how to configure CMSSW from python
• translating cfg files to python

Misc.

• AOD data content table
• creating an user code area in CVS
• my CMSSW user code in CVS
• graceful kill (i.e. finish processing the current event and close output file) of a CMSSW job: kill -USR2 ... (see also here)
• description of ΔR matching tools in CMSSW
• CMS Top quark physics group page

-- AndreHolzner - 10 Jul 2008