GDE ILC Reference Design Report including short physics chapter

1
Documents to be produced by the end of 2006
Klaus Moenig Ties Behnke

• GDE ILC Reference Design Report (including short
  physics chapter)
• Detector Concepts Report (including longer
  physics chapter)
• timescale Bangalore (March) presentation and
  discussion of outline
• Vancouver (July) detailed discussion
  with the community
• Valencia (November) presentation of
  final draft
• DCR performance chapter look at small number of
  selected benchmarks,
• distribute the reactions among the concepts ?
  avoid entering into competition at this stage
• currently no solid and understood full
  simulation, although a lot of recent progress
• e.g. when running current
  reconstruction on GEANT4 detector simulation,
  goal of
• DE/E 30/sqrt(E) not yet achieved ?
  rely on fast simulation in the mean time
• T Behnke unlikely to be achieved on the
  timescale of DCR
• ? need to be careful not to
  draw wrong conclusions

2
Recent developments (LDC software) I
Frank Gaede

• LCIO plans check usability of data model, e.g.
  track parameters
• need to make LCIO more convenient
  and efficient, e.g. allow requesting all
• tracks for one MCParticle
• Marlin (Modular Analysis Reconstruction for
  the LINear collider)
• software modules called processors, each has
  set of input parameters,
• order of processors is defined via XML
  steering file ? easy to exchange modules without
• recompiling, can run the same processor with
  different parameter set in one job
• Strong encouragement to register with the GRID
  (http//lcg-registrar.cern.ch)
• (example DESY 10 outdated ILC-dedicated
  machines, GRID 300, beyond Europe few 1000)
• GRID has already been used to produce event
  samples, accessible on the web
• 450 k events at 500 GeV Z0, uds, cc, bb,
  ttbar, WW, ZH

3
Recent developments (LDC software) II
Steve Aplin

• Tracking status (based on LEP code previously in
  BRAHMS)
• TPC pattern recognition modified ALEPH code,
  final track fit from DELPHI Kalman-filter
• VTX- SIT-hits treated as individual track
  element cov matrix, included in trk
  matching/fitting
• new Marlin processor to suppress low energy
  curlers, physics impact needs to be determined
• kink finding not yet included
• efficiency studies to be done soon, initial
  studies gt 93, but track splitting still a
  problem
• Particle flow algorithms
• PFA performance at high energies unacceptable
  due to deficiencies in other parts of rec code
• missing kink-finding gave rise to large tails
  in E(rec) E(true) distribution
• MC true information only available at particle
  origin, not at point of kink, and currently no
• place to store this information in LCIO
• errors from clustering and from track-cluster
  matching due to differences in calibration
• gaps between subdetectors, missing in geometry
  available for reconstruction ? split clusters
• ? urgent need to solve some technical issues

Predag Krstonosic
4
Reaction to LCFI presentations

• Vertexing flavour tagging highlighted as
  missing in the current reconstruction software
• in a number of presentations
• response to our talks was therefore very
  positive
• Frank Gaede emphasised need to make our code
  available in the form of
• Marlin processor(s) in order to integrate it
  into the LDC software framework
• Ties Behnke gave advice to think about which
  changes will need to be made in
• LCIO classes to allow output from our code to
  be stored well in advance of the release
• Mark Thomsons summary talk
• vertex reconstruction sophisticated code under
  development real progress
• vital for many physics studies
• high priority to get this in Marlin
• promised by summer