Minutes of 23 March 2016 Meeting of Geant4 Technical Forum
Editor: A. Morsch, 25 March 2016

- Starting at 16:10

Highlights of recent patch releases - non-physics part
======================================================
+ see slides



Highlights of recent patch releases - physics part
==================================================
+ see slides

- Clarification on the validation results:

 Shower shape differences are expected if W or Pb used in calorimeter.
 They are slightly narrower, i.e. closer to the HP option but slightly
 worse with respect to data.


2016 work plan (non-physics part) and supporting platforms
==========================================================
+ see slides

- explanations

 - scaled shape constructor: possibility to scale shapes in dimension
 for a single shape
 for example tube -> elliptical tube or sphere -> ellipsoid
 - Mongo data base: cross-platform data base to store ntuple-like
 analysis objects
 - material scanner: shooting geantinos to scan material properties


2016 work plan (physics part)
=============================
+ see slides

- Request: description of issues sometimes cryptical; where possible give
 link to JIRA which has a more detailed description of tasks.


Open Requirements
=================
+ see slides

- Vladimir: 3301, 3701 and 3602 are not relevant anymore for CMS


**** New Requirements

ATLAS
=====
+ see slides

- slide 3 in future, Geant4 tests with simplified geometry should
 include variation of angle of incidence (now only 90 deg)
- slide 4 understanding low-pT JES change between Run1 and Run2
 simulations has highest priority.
 Cross-section change in Bertini ? ATLAS will have to test first
 how v10 performs.
- slide 5 issues with very small steps (fm), observed for all energies
 but separated by particle type, always for charged particles,
 independent of magnetic field, for long and short tracks,
 no particular geometry
- slide 6, 2nd bullet: concerns in general all cross-sections not
 only hadronic


CMS
===
+ see slides

- slide 2: what is interest in gcc60 ?
 More natural usage of co-processors


ALICE
=====
+ see slides 

- computing time g4/g3 now ~ 1.6
- transition to g3 depends on ongoing validation tests.
 Still problems with TRD detector. Better performance with G3 wrt
 G4 PAI + retuning.