Event Filter studies

1
Event Filter studies

• Giovanni Siragusa
• INFN-Univ. of Lecce
• for the TrigMoore Working Group

PESA Performance and Validation November 27, 2006
2
Outline

• Software status
• Samples and releases
• Performance studies to validate samples and
  reconstruction suite
• Resolutions
• Efficiencies
• Consistency checks
• Summary and plans

3
Software status

• Vertical slice configured for trigger menus at
  900GeV in addition to standard configuration
• RoI dimension can be increased according to
  b-physics requirements
• AAntuple ready and working (presently under test)
• For the first time TrigMoore uses RoiDescriptor
  in the endcaps when phi and eta are not given by
  the MuonFeature
• AOD and ESD are produced on an event basis
• as soon as possible (in 12.0.4?) they will allow
  rerunning of the Steering (HypoAlgos) via key2key

4
Samples and Releases
Single muon pT 8, 20 GeV/c CSC production with
Athena 11.0.42 (Napoli Tier2 CERN) DC3-02
Layout Top events /castor/cern.ch/atlas/csc/valip
rod/sampleA/testIdeal_07/005200.T1_McAtNlo_Jimmy/d
igit/v12000201/testIdeal_07.005200.T1_McAtNlo_Jimm
y.digit.RDO.v12000201_tid002760._00002.pool.root.1

• Reconstruction with Athena 12.0.X, TrigMoore
  seeded by LVL2 result available over barrel
  endcaps
• Moore for MS track reconstruction.
• MuId StandAlone for MS track extrapolation.
  (Parametrization of dE/dx in the calorimeters)
• MuId Combined for MS and ID tracks matching.
• The Event Filter is now, for the first time,
  seeded by MuComb (last step of LVL2 combining MS
  and ID tracks). Efficiencies wrt LVL2 are now
  intended wrt MuComb (and not muFast).
• Hypo Algos running by default at LVL2 and EF
  level

Preliminary results low statistics, many fixes
are on going
5
pT rec / pT gen

• Red MuidSA, for tracks successfully fitted
• Pink MuidCB, for tracks successfully fitted

No relevant bias within statistical errors
s(pT)/pT vs h
pT20 GeV
Sigma
Mean
h
h
Compatible with release 12.0.3 results, more
statistics should be run for clearer evidence
6
Muid pT resolutions
pT20 GeV
pT8 GeV
Completely compatible with release 12.0.3 results
MuidSA
MuidCB
7
pT resolution vs pT
h lt 2.5
Release 12.0.3
Combined reconstruction by MuidCB (including
Inner Detector) improves reconstruction
performance by MuidSA (Muon Spectrometer only).
8
Moore-Muid Efficiency
Efficiency are defined wrt LVL2
pT20 GeV
Moore
MuidSA
9
Summary of efficiencies
Efficiencies as expected Problems with
MuidCombined efficiency in 12.0.3 disappeared
10
Consistency checks

• Muon slice works smoothly on top events
• /castor/cern.ch/grid/atlas/dq2/calib1_csc11//calib
  1_csc11.007234.singlepart_mu200.digit.RDO.v1200301
  //calib1_csc11.007234.singlepart_mu200.digit.RDO.v
  12000301_tid003123._00002.pool.root
• Produced AOD, ESD and CBNT ntuple.
• Quick checks were done on 100 events

11
Consistency checks-2
pTrecCB/pTtrue
pTrecCB/pTtrue
if pTtrue gt10 GeV
if pTtrue gt4 GeV
pTrecCB/pTtrue
AOD
if pTtrue gt20 GeV
px
12
Short-medium term plans/involvements
M.Primavera

• Muon EF for cosmics As soon as the Trigmoore
  migration in the new EDM will be completed, the
  Lecce group working on Muon EF will provide and
  test a specific implementation of Trigmoore for
  cosmic muons by adapting in the Trigmoore
  environment the offline (new EDM) package Modular
  MuonReconstruction for cosmics (Nikhef)Muon
  Data validation for CSC Data sets SW release
  validation (e.g. nightly tests) on single muons
  and physics samples (Z-gt??, H-gt4?). G. Siragusa
  appointed in the Lecce group to implement the
  tools and perform the job. Preliminary discussion
  on work organization/suggestions from M. Wielers,
  responsible of Electron data validation. Data
  quality monitoringData quality assessment and
  establishing for EF muons (contact people in
  Lecce M. Primavera). Preliminary discussions
  with R. McPherson on organization, infrastructure
  needed, etc.