MiniAOD Studies

Repository for MiniAOD Studies for CMS BPAG.

Track Details & Covariance Matrix

A first approach to understand if miniAOD may be used for "all" BPAG analysis would be to act on the PFPackedCandidateProducer lowering the pT thresholds:

minPtForTrackProperties that selects the minimum threhsold so that a track is stored (i.e. with its covariance matrix) with full precision (here) and set by default to 0.95 (here)
the minimun for a track to be stored with a reduced precision covariance matrix, hard coded to be 0.5 (here)

In order to make also the latter configurable at run time a new variable minPtForLowQualityTrackProperties is defined in PATPackedCandidateProducer to select a threshold different from 0.5 GeV. All the (few) changes are stored in a branch of CMSSW_11_2_X: miniaodsizes.

To be able to run with it merge the topic in your local repo

cmsrel CMSSW_11_2_0_pre6
cd CMSSW_11_2_0_pre6/src/
git cms-init
git cms-merge-topic adrianodee:miniaodsizes
scram b -j 8

and then you will be able to run the MINIAOD step (such as the ones stored in configs/ directory) and adding a couple of lines as

process.packedPFCandidates.minPtForTrackProperties = x
process.packedPFCandidates.minPtForLowQualityTrackProperties = cms.double(0.1)

will allow you to tune the reduced precion covariance matrix pT threhsold. For a preliminary study the configuration that have been taken into account are

	minPtForTrackProperties	minPtForLowQualityTrackProperties
Original	0.95 GeV	0.50 GeV
v0	0.10 GeV	inf (*)
v1	0.50 GeV	0.10 GeV
v2	0.30 GeV	0.10 GeV
v3	0.95 GeV	0.00 GeV
v4	0.00 GeV	inf (*)
v5	0.50 GeV	0.25 GeV

(*)in this case this threshold doesn't have any effect

and you can find the results for TTbar (in update...) events with <PU> = 40,45,50,55,60,65 in the notebook Sizes. Here some examples of the results.

Packed Candidates Collection Avg Size

MiniAOD Relative (to orginal) File Size

Sample locations

All the samples have been produced with 11_2_0_pre6 and GT 'auto:phase1_2021_realistic’ (i.e. 112X_mcRun3_2021_realistic_v7)

	location	events	config
dstar (pT>3.9)	/eos/cms/store/group/phys_bphys/miniaodstudies/dstar/	~90k	DStarToD0Pi_D0KPi_pT3p9
dstar (pT>0.5)	/eos/cms/store/group/phys_bphys/miniaodstudies/dstar_lowpt/	~50k	DStarToD0Pi_D0Kpi_pT0p5
bsjpsiphi	/eos/cms/store/group/phys_bphys/miniaodstudies/bsjpsiphi/	~50k	BsToJpsiPhi
qcd_muon	/eos/cms/store/group/phys_bphys/miniaodstudies/qcd_muon/	~60k	QCD
bupsik	/eos/cms/store/group/phys_bphys/miniaodstudies/bupsik/	~30k	BuPsiK
xilambda	/eos/cms/store/group/phys_bphys/miniaodstudies/xi/	~50k	Xi

Checking the LUT

In order to check the LUT for the covariance you may merge https://github.com/AdrianoDee/cmssw/tree/miniaodbph branch in your CMSSW

cmsrel CMSSW_11_2_0_pre6
cd CMSSW_11_2_0_pre6/src/
git cms-init
git cms-merge-topic adrianodee:miniaodbph
git cms-checkdeps -a 
scram b -j 8

This would allow you to access the covariance matrix (for the pseudoTrack of the PackedCandidate) even when hasTrackDetails() is false and the track is attached to the candidate. In order to get the track reference in this setup check before the covarianceVersion to be ==1 (or simply >0) with the method covarianceVersion()from here and then get the reco::Track with the method pseudoTrack() from here.

slezki / miniaodstudies