Run II Reconstruction Program Status - Version p08

General

• Cleanup of error messages (use ErrorLogger).
• Addition of general statistics package (recostat).
• Support for Raw Data Chunk produced by p07 Monte Carlo.
• New versioning implementation for detector geometries.
• Use geometry database system, instead of "in memory" mode.

Detectors

• SMT
• CFT
  • Support for distortions to describe as-built geometry.
• CPS
  • New method for finding preshower energy in cone (eCone).
  • Basic handling of saturated strips.
  • Tuned parameters.
• CAL
  • Use pmc02 derived, eta-indep, cal-only weights (CPS not included).
  • Fix problem with handling zero energy towers.
• ICD
• FPS
• MUO
• FPD (not yet in RECO)
• L0 (not yet in RECO)

Algorithms

• Global Tracking
  • Turn on overlap tracking.
  • Use non-uniform magnetic field for track finding in the overlap region, and for refitting in overlap and central region.
  • Better accounting for CFT cluster errors in an attempt to solve the CFT efficiency drop-off with increasing number of background events.
  • Better (more efficient) SMT extension paths.
  • Numerous speed-ups and optimizations.
    • CFT central tracking has been split into separate axial and stero steps.
    • SMT geometry has been optimized (speeds up refit).
    • Various computer science optimizations (static linear algebra, local reference count smart pointers, cluster containers).
• Vertexing
  • Primary vertexing
    • Single track vertexing
    • Better selection
      • Minimum track PT-sum of 5 GeV, to remove spurious vertices
      • Ability to choose maximum multiplicity vertex as PV
      • Better min. bias separation from hard scatter vtx
  • Secondary vertexing
    • Kalman vertexing
      • Track clustering
      • x40 speed-up
    • Secondary vertexing
      • Better seeds
      • Speed-up

Object ID

• EM ID
  • Implement "road method" in SEMRECO (not yet used to identify electrons)
  • Fix bug in CellNN that generated zero energy clusters
  • Protect against crashes from zero energy clusters.
• Muon ID
  • New chunk: Muon:SegmentChunk
  • New segment object:
    • Knows its position, direction, and errors
    • Knows matching scintillator time
    • Methods for retrieving PDT, MDT and MSC hits
    • Method for a linear propagation to facilitate matching with calorimeter and central track.
    • Not yet used in muon id (scheduled for p09)
  • New segment finding algorithm: LinkedList (now default)
  • Local track fitting upgraded (WAMUS and FAMUS)
• Tau ID
  • New flag word in tau object (FLAG = 1 indicates tau matched to electron)
  • New EM fraction calculation using 0.3 cone (not cut on yet)
  • New profile cut (> 0.25)
  • Retrained H-matrices on events from p05.00.04
  • Added tau MC info and H-matrix info to reco_analyze ntuple.
• Jets / MET
  • Minor fixes to jet algorithms
    • Change longitudinal matching of floor clusters in CellNN to handle EC and CC similarly.
    • Fix inconsistencies in handling "parton" and "partons".
  • Calculate MET in eta rings with vertex = (0,0,0), to improve ability to handle revertexing
• Heavy Flavor ID
  • bcJet chunk now persistent (version 1.0)
  • bcjet_analyze part of reco_analyze
  • Five taggers:
    • eTagAlg (semreco)
    • muTagAlg
    • SecVtxAlg
    • JetImpAlg
    • LikelihoodAlg

	Symptom	Status
	No high PT tracks found in GTR	fixed in p08.01.00
	Noisy bcjet_reco.	fixed in p08.01.00
	Farm script configuration problems.	fixed in p08.01.00

• Fix geometry system to load correct CFT geometry. Cured problem of finding no high PT tracks.
• Suppress some bcjet_reco output.
• Fix farm scripts.

	Symptom	Status
	Floating point exception in GTR on Irix.	fixed in p08.02.00
	No default geometry for real data runs.	fixed in p08.02.00
	Geometry system does not yet support real data files.	fixed in p08.02.00
	Noisy bcjet_reco.	fixed in p08.02.00
	Bug in bcjet_reco:eTagAlg. Crashes on events already processed.	fix scheduled for p08.03.00
	Rare crash when processing p05 MC events. /D0/dist/releases/p08.01.00/include/vertexmatch/Match2.c:88: failed assertion `pqold->is_valid()'	fixed
	Unpacking code uses a pattern that does not work with optimized builds. Causes problems when processing raw data chunks using optimized builds. The following pattern does not work: return ((signed short int)(*_buffer & 0xFFFF) >> 4);	fixed
	Wrong HMatrix file (hmatH41 uses hmatL41.dat)	fixed
	Cal weights are not automatically selected for p07 / p08 MC events.	fixed
	CFT geometry loads every event, causing slower event processing.	fixed
	Many debug messages printed out when processing raw data.	fixed
	Error messages clog .out file when processing raw data - need to use ErrorLogger.	fixed
	New MET ntuple words not yet installed.	fixed

• Fix floating point exception in GTR on Irix.
• Add default geometry for real data.
• Suppress more bcjet_reco output.

	Symptom	Status
	Bug in bcjet_reco:eTagAlg. Crashes on events already processed.	fix scheduled for p08.03.00
	Crash in recostat when processing real data (partial detector / cosmic runs): %ERLOG-e UNKNOWN: Can't find type information for object. %ERLOG-e UNKNOWN: Typeinfo name: bcjetid::bcJetChunk Occurs when there are not valid GTrack chunks.	fixed
	Rare crash when processing p05 MC events. /D0/dist/releases/p08.01.00/include/vertexmatch/Match2.c:88: failed assertion `pqold->is_valid()'	fixed
	Rare crash when processing p05 MC events. /D0/dist/packages/cft_cyl2D/v00-01-85/src/ClusterFilterR0_Ptmin_Zchi_2Hit.cpp: 532: failed assertion `x_bound1 > -almostzero && y_bound1 > -almostzero'	fixed
	Unpacking code uses a pattern that does not work with optimized builds. Causes problems when processing raw data chunks using optimized builds. The following pattern does not work: return ((signed short int)(*_buffer & 0xFFFF) >> 4);	fixed
	Crash when processing real data (partial detector / cosmic runs): ZMthrow was issued at line 34 /D0/dist/packages/l3base/v00-05-03/src/L3Crate.cpp	fixed
	Crash when processing real data (partial detector / cosmic runs): /D0/dist/packages/muon_geometry/v01-03-02/base/MuoGeometer.cpp:388: failed assertion `cc != 0'	fixed
	Crash when processing real data (partial detector / cosmic runs): /D0/dist/packages/muon_geometry/v01-03-02/base/MuoGeometer.cpp:381: failed assertion `wmsc != 0'	fixed
	Rare crash in kt_jets::makePreKtClusters.	fixed
	Wrong HMatrix file (hmatH41 uses hmatL41.dat)	fixed
	CFT geometry loads every event, causing slower event processing.	fixed
	Many debug messages printed out when processing raw data.	fixed
	Error messages clog .out file when processing raw data - need to use ErrorLogger.	fixed
	Cal weights are not automatically selected for p07 / p08 MC events.	scheduled for future pass release
	New MET ntuple words not yet installed.	fixed
	New parameters for A / BC segment matching in muon id.	fixed

Problems reported related to reco_analyze

	Symptom	Status
	Crash on real data (partial detector / cosmic runs) in vertex_analyze: Program received signal SIGSEGV, Segmentation fault. L4vec::px (this=0x0) at /D0/dist/releases/p08.02.00/include/mcpp/L4vec.hpp:87 87 return _px;	fixed
	Crash on real data (partial detector / cosmic runs) in vertex_analyze: Program received signal SIGSEGV, Segmentation fault. /D0/dist/packages/geometry_management/p08-01-00/src/geometry_management.cpp:114	fixed

This pass release fixes problems with the maxopt build. Users are encouraged to use the maxopt build. It runs approximately x2 faster.

	Symptom	Status
	p08 reco does not support processing p08 Monte Carlo events.	scheduled for p09
	p08 reco has limited support for processing raw data.	scheduled for p09
	SAM executable crashes trying to include RecoStat package "Unable to make package of type RecoStat with name recostat"	fix scheduled for p08.11.00
	Can't reprocess previously reco'd p05 events.	fix scheduled for future release

• SMT
• CFT
• CPS
• FPS
• CAL
• ICD
• MUO
• FPD
• L0

• Tracking
• Vertexing
• EM ID
• Muon ID
• Tau ID
• Jets/MET
• Heavy Flavor ID