Minutes of 16 November 2010 Meeting of Geant4 Technical Forum

Editor: A. Morsch, 20 November 2010


- Starting at 16:40


Updates on open requirements
============================
  + See slides
  Recent Requirements
  + #2801 Light anti-Ion Transport: Good progress on recent request from ALICE 
  + #2703 Accepted and solution proposed, will be included in plans for 2011 
  + #2701 G. Corti: no feedback from LHCb yet 
  Recent Updates 
  + #2501 CMS (S. Banerjee) confirms: request not longer an issue. 
 
Geant4 9.4 release and latest patches
=====================================

Updates on Electromagnetic Physics
=================================
  + See slides
  Discussion on slide 20:
  John: Option0 is default. Urban model used for e+e- WentzelVI for muon
        transport.
  No effect on overall computing time (typical LHC event); 10% for pure muon
  transport.

Updates on Hadronic Physics 
===========================
  + See slides
  - To slide 4: Longitudinal showershapes have improved significantly between
    7.1 and 9.3 but still work to do.
  - To slide 5: J. Apostolakis: Following the improvements made on the FTF
    model, QGS will be revisited and revised.
    CMS is concerned about performance. D. Wright: no significant cost of
    performance.
 

Users' presentations
====================
LHCb
==== 
  + See slides  
  - To slide 7:
   "G4 experts (V.I.) to set up a customized EM PL with correct treatment of
    MCS for all particles"
    Comment V.I: Technically possible but no promise
  - To slide 8:
    Vladimir: Variation of calo calibration parameters with G4 release not
    observed in simplified test set-ups. 
    Have trending since G4 7.0 and variation < 1% or no change at all.
    Needs better understanding. 
    Gloria: delta-rays as reason excluded.
  - To slide 11:
    Need expert's help to understand hanging jobs.

ATLAS
=====
  + See slides
  - To slide 5: Makoto: Need better understanding of the observed 100% CPU
    overhead in cavern background simulations. 
  - To slide 7: observed difference not a G4 issue.
    John: concerning memory scaling between 32/64 bits, recommendation to talk
          to CMS

CMS
===     
  + See slides
  - To slide 5: Makoto: G4 9.3 until when ?
    CMS started validation on 9.4 candidate. Move to G4 9.4 by the end of the
    year, except, if there are severe problems. 
    After that, they need 2 years stability.
  
ALICE
=====
  + See slides
  - First central GRID productions with G4. Full detector simulation of 7 TeV
    data and comparisons with 
    G3 and data. Differences seen but not yet a Geant4 issue, still
    investigating.
    John: Lower secondary particle yields known issue for QGSP.

Concluding remarks
==================
John: 
- Experiments interested to move to 9.4.
- Release will be on time

Gabriele:
- We make candidates available as they come out
- Let us know if you are interested and provide feedback !

Closing meeting at 19:00.