Production cache 12.0.6.4

• /afs/cern.ch/atlas/software/releases/AtlasProduction/12.0.6.4

• Manuel Gallas

• CMT v1r19
• SLC3
• gcc3.2.3

• List of Tags
• Tag differences compared to previous cache (12.0.6.3)

• Distribution kit ready

As a pacman kit:
(you should setup first 12.0.6 release: source ./cmtsite/setup.sh -tag=12.0.6)

pacman -get http://cern.ch/atlas-computing/links/kitsDirectory/Production/cache:AtlasProduction_12_0_6_4_i686_slc3_gcc323_opt
which will be mirrored to classis (INFN):
pacman -get http://classis01.roma1.infn.it/pacman/Production/cache:AtlasProduction_12_0_6_4_i686_slc3_gcc323_opt

(1)
TileCalorimeter/TileConditions-00-03-42
TileCalorimeter/TileSimEvent-00-01-19
TileCalorimeter/TileSimAlgs-00-02-27
- Justification:
New TileConditions tag fixes bug with MBTS cells -
with previous version twice more MBTS cells (64 instead of 32)
with duplicated IDs were created in output container.

New TIleSimAlgs and TileSimEvent allow to simulate
photostatistics in TileTTL1 (and therefore in MBTS trigger output),
similar to what was provided  for normal TileCal cells.

(2) 
Generators/DC3_joboptions/DC3_joboptions-00-01-33
- Justification:
 just joboption updates and efficiency fixes as usual

(3)
Simulation/Digitization/Digitization-00-08-31-03
- Justification:
The only change wrto the release tag is a property setting in 
ConfigPileUpEventLoopMgr.py ToolSvc.CavernCache.ReadDownscaleFactor=50
This should mitigate the non-physical hit distributions shown by 
at recent Higgs meetings.

(4) 
Simulation/SimuJobTransforms/SimuJobTransforms-00-07-10-09
- Justification:
It adds some jobOptions frangments requested by the Bphys group

(5)   
Simulation/Atlfast/AtlfastUtils/AtlfastUtils-00-03-08
- Justification:
Fix to accommodate recent changes in the truth block from Simulation

(6) 
Tools/PyJobTransformsCore-00-06-47
- Justification:
 ignore the errors:
     MuonRdoToMuonDigit                       ,  ERROR ElemetID not found          
     TgcRdoToTgcPrepData                      ,  ERROR ElemetID not found 
 found in the stream test 

(7) 
PhysicsAnalysis/AnalysisCommon/UserAnalysis-00-09-10-16
PhysicsAnalysis/AnalysisCommon/UserAnalysisUtils-00-01-01-09
PhysicsAnalysis/AnalysisCommon/UserAnalysisEvent-00-01-06-03
PhysicsAnalysis/AnalysisCommon/ParticleBuilderOptions-00-00-28-09
- Justification: 
User wants more information on MissingET 

(8)
PhysicsAnalysis/EventViewBuilder/EventViewConfiguration-00-00-49
- Justification: 
Various Minor HPTV related fixes

(9)
PhysicsAnalysis/SUSYPhys/SUSYView-00-00-19
- Justification: 
HPTV-Base SUSYView for SUSY WG DPD Production

(10)
PhysicsAnalysis/HighPtPhys/HighPtView-00-00-27
- Justification: 
Various fixes relating to selection criteria and ntuple content

(11)
Reconstruction/RecJobTransforms-00-05-41
- Justification: 
To fix crash in production related to TAGoutput file name(bug 25707).
AOD 1mm range cut correction 

(12) 
PhysicsAnalysis/ElectronPhotonID/ElectronPhotonAlgs-00-03-03
PhysicsAnalysis/JetMissingEtID/JetMissingEtAlgs-00-01-14 
- Justification: 
AOD 1mm range cut correction

(13)
PhysicsAnalysis/EventViewBuilder/EventViewUserData-00-00-64
PhysicsAnalysis/EventViewBuilder/EventViewInserters-00-00-49 
- Justification: Truth IsGenStable Fix:

(14) 
MuonSpectrometer/MuonReconstruction/MuonSegmentMakers/MuonSegmentMakerAlgs/MooSegmentMakerAlgs-00-00-22-01
  Justification: 
Bug reported by the central CSC production, see http://savannah.cern.ch/bugs/?25008 which is related to http://savannah.cern.ch/bugs/?24417. The problem is that for the CSC J7 sample, a crash in the reconstruction was traced to Moore spending too much time due to a bug.
It was tested by reconstructing 9350 Z->mumu events with the official 12.0.6 and with the modified 12.0.6 + MooSegmentMakerAlgs-00-00-22-1. The performance results can be summarized as follows:
- number of Moore tracks = 14303 (12.0.6) vs. 14306 (12.0.6+tag)
- number of MuidCb tracks = 13405 (12.0.6) vs. 13403 (12.0.6+tag)
- with the added tag there are more tracks with 7 or 8 CSC eta hits or with 2-MDT layers in the inner station (i.e. more hits in the overlap regions)
- Moore 1/pT resolution = 3.06% (12.0.6) vs. 3.04% (12.0.6+tag)
  So there is little change for "standard" physics events whereas the very large processing time is cured for high-energy jet events that leak into the Muon Spectrometer. It is expected the bug will also show up in the samples to be processed with pileup and cavern background.

(15) 
PhysicsAnalysis/AnalysisCommon/ParticleBuilderOptions-00-00-28-10
- Justification: 
  added corrected MissingETs to SAN 

• More info in the ATLAS PyJobTransforms Wiki