RECO Status Report - July, 2002

Status of releases:

• p11.09.00 - Running on data farms.

  Known problems:

  • reco_analyze cannot be run on p10 reco files (crash in bcjet).
  • Monte Carlo production fails in reco_analyze due to re-occuring bug related to network access.

• p11.10.00 - Will be installed soon.

  New features:

  • Fix ICD mapping.
  • Fix calorimeter non-linear corrections for Monte Carlo.
  • Fix CPS reco_analyze package and re-enable.
  • Fix bug causing connections to db to crash.
  • Fix thumbnail bugs for EM reco.
  • Fix MET block in reco_analyze.
  • Remove some noisy printout from CFT.

• p12
  • p12.01.00 being built.
  • RECO runs. RECO_ANALYZE fails.
  • Possible problem with memory consumption.
  • Still under investigation.

Recent progress:

• With two minor fixes to d0reco and calibration_manager, a version of p11.10.00 RECO successfully used the SMT calibration database.

Major outstanding problems:

• With the lowering of the calorimeter thresholds to 1.5 sigma, the RECO processing time has gone up substantially. In addition, the size of the RECO_ANALYZE root tuples has expanded. Currently under consideration: optimizing unpacking code, applying offline thresholds after non-linear corrections are made, optimizing size of root tuples.
• Without access to the calibration databases, official running of RECO requires SMT and CFT flat files, which are available in the head versions of CVS packages. Non-farm users have been unaware of this requirement. Currently under consideration: improve documentation, automate access with d0tools.
• RECO_ANALYZE crashes on Monte Carlo farms in p11, due to what appears to be a reoccuring bug related to connections to the network. Under investigation

Ongoing projects:

• Tracking algorithms for p13 - A committee was recently appointed to collect data concerning the currently available tracking algorithms and make recommendations concerning which algorithms should be run in p13.
• RECO Monitoring tasks - Currently four people have volunteered and are working on the two advertised RECO monitoring projects.
• Chunk documentation - Currently one person has volunteered.