Simulation Release Notes for the 3_1 cycle

Introductory note

The current release notes refers to the production setup deployed as of 3_1_1. Several other capabilities are contained in the release, and needs to be further integrated and/or refined:

  • Alternative calorimeter simulation using GFlash
  • Simulation of the forward region currently not described, e.g. CASTOR and ZDC, including the beam transport down to ZDC by using Hector
  • tool to merge real data and simulated events (DataMixingModule)

The description of these components will be added in case they will be integrated in further versions.

Known problems

  • The RPC geometry in the negative side has presumably a problem: the position of the SimHits is correct, but the momentumAtEntry originating the SimHits is returned with a swapped value. This has no evident impact at reconstruction level, and therefore has to be considered as a known feature to be fixed in future (4_X_Y) cycles provided strict backward compatibility in the geometry description can be released there.

Updates in 31X

3_1_2

  • Fix in StackingAction (SimG4Core/Application): decay products inside the beam pipe were not always saved, since the beam pipe was partitioned in several volumes with copies (positive and negative sides), and the tests to decide whether to save or not the particle for a given volume were performed on the Physical and not on the Logical Geant4 volume, e.g. considering only one of the copies.
  • Fix in ECalSD (SimG4CMS/!Calo): protection to avoid that the argument of a log() can be zero (null dedx in a G4Step)
  • Fix in SimMuon/!RPCDigitizer: random number generator moved from RandPoissonQ to RandPoisson to avoid FPE when the mean of the Poisson distribution is very small (effectively hiding a lack of protection inside the CLHEP class).

3_1_3

  • Fix in TrackingVertex for non zero pileup events (correct BX association)
  • Fix of memory leak in GFlash NB: ECAL digitization adjustment to be fixed as in CMSSW_3_3_0_pre6
  • MixingModule pre-merging for multiple external sources
  • Configurations and cmsDriver settings for DataMixingModule

Geometry

  • Preshower back in default configuration
  • Optional extended configuration to include CASTOR (empty box in default configuration), ZDC and possibly in future Totem and FP420
  • Beam pipe split in several pieces to optimize the navigation

Tracker Pixel

  • added the description of cables and PP0

Geant4

Geant4: new version 4.2.patch01 (see its release notes). Physics list: QGSP_BERT_EML . A variant of the previously used QGSP_BERT_EMV, differing in the treatment of the electromagnetic part (use of the ApplyCut option for performance improvements). Physics performances are very similar for production cuts not exceeding 1 mm: the tracker passive material ones and the ones of the muon chamber walls have been adapted accordingly. The internal handling of simulated tracks and calorimetric hits has been move on a "per primary track" basis, allowing the intermediate storage of needed objects, cleaning of dropped ones and re-usage of freed memory. Introduced cuts in the stacking and stepping actions to discard tracks whose time is exceeding a given threshold separately for each G4Region: used to suppress the simulation in non interesting times of long traveling particles (especially neutrons) where needed, allowing the correct treatment of the forthcoming simulation of forward calorimeters (where the time of flight to reach them is comparable with the cut in the central part of the detector).

The simulation of the forward region (e.g. ZDC) is optionally available (needs GeometryExtended_cff.py and SimExtended_cff.py). It relies on the transport of particles down the beam pipe by the Hector Hector external package, through the module LHCTransport. In CMSSW a copy of the generated event is produced by the module LHCTransport, with the particles transported attached as outgoing from a vertex at the position at the end of the transport and this copy is used as input by the module g4SimHits (e.g. Geant4).

Pile Up simulation (MixingModule)

The CrossingFrame has become a non persistent collection, and internally it uses pointers to the pileup containers to reduce the memory used. Only the time is a separate variable. The change is transparent to the end user.

Digitization

Tracker Pixels

  • Pixel threshold smearing
  • Pixel geometry (cables to PP0)
  • Pixel DB access (incl. dead modules)
  • New values for Pixel thresholds and injected signal amplitude to electron conversion
  • the ReadoutNoiseInElectron has been set to 350e (was 500e in 22X).

Tracker Strips

  • New parametrization of the cross-talk per geometry, and refined parameters for TIB and TOB, using measured value from CRAFT (in peak mode).
  • New parametrization of the pulse shape (both in peak and deconvolution mode) to better take into account the tails.
  • Introduced a mean inefficiency per module (set to 0 by default).
  • New default (more realistic) parameters for the digitizer.
  • Preparation of config files for operation in deconvolution mode.

ECAL

  • noise constant in ADC counts (was constant in energy in 22X)
  • simulation of dead channels done at the reconstruction level, via a configuration flag
  • preshower fully integrated in the official setup

HCAL

  • time smearing added

Muon DT

Muon CSC

Muon RPC

  • in CMSSW_2_X only mu+/mu- SimHits are digitized, while in CMSSW_3_1_X all but electrons are digitized
  • RPCSynchronizer - signal propagation time is taken into account in CMSSW_3_1_X
  • safer parameter reading

Contacts

Review status

Reviewer/Editor and Date (copy from screen) Comments
FabioCossutti - 9 Jul 2009 initial skeleton
KatiLassilaPerini - 23 Jan 2007 created template page

Responsible: ResponsibleIndividual
Last reviewed by: Most recent reviewer

Edit | Attach | Watch | Print version | History: r8 < r7 < r6 < r5 < r4 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r8 - 2009-10-01 - FabioCossutti
 
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    CMSPublic All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback