Minutes of the Geant4 Technical Steering Board Meeting 6 May 2004

Submitted by Anonymous (not verified) on

Version 1.0

$Date: 2016/02/08 19:15:25 $



Present:

CERN: John Apostolakis (chairman), Gabriele Cosmo, Andrea Dell'Acqua, Gunter Folger, Vladimir Ivantchenko, Fred Jones (on behalf of Peter Gumplinger), Maria Grazia Pia, Sergei Sadilov and Hans-Peter Wellisch.

KEK: Katsuya Amako, Takashi Sasaki, Koichi Murakami

Naruto: Hajime Yoshida

SLAC: Makoto Asai, Dennis Wright

ESA/ETEC (by telephone): Giovanni Santin

Madrid (by telephone): Pedro Arce

Excused:

Joseph Perl

Minutes: G.Cosmo - 10 May 2004

  • NOTE: the meeting starts with 30 minutes of delay due to technical problems with ESnet connections with SLAC.

Initial items

  • Minutes of previous meetings were approved.
ACTIONS
  • Closed actions: 18.5, 27.2, 27.5, 27.8, 27.12, 27.15.
  • 18.4: changes in Bonsai implemented. Will be made available together with other improvements before mid-May 2004.
  • 18.12: JA reported that no volunteers have been found. MGP maybe know some people who could take the responsibility. Maintenance of the training-toolkit should also be seen in relation with the material eventually distributed in the CD.
  • 22.4: a report is expected for the next TSB in July 2004.
  • 24.3: a publication policy should be expressed in the CA. There's need of a policy for publication and registration: a possibility is to put in place an e-mail reminder for notifying of papers published (or to be published) at every quarter. Improvement and update of the existing structure and web pages is needed, as well as keep up to date the existing AFS repositories for paper sources.
  • 27.2: the requirement has been inserted in the UR Keeper. The assignee must be corrected to coordinator of Fast-Simulation & Parameterisations. Action is closed.
  • 27.6: DW reports that the scope is to have a companion to the hadronic physics-lists (where appropriate to use, what models/ cross-sections cowork, ...); users require this kind of explanations. The effort required for achieving this is estimated in 1 person working for 1 month full time.
  • 27.7: SS reports all will be in place before mid-May 2004.
  • 27.13: the reorganisation is related to filling the the gap between physics-reference manual and the rest of the documentation, and to the restructuring of the Toolkit Guide.
  • 27.14: MGP reports that work is in progress for a 'tentative' volume for EM physics. A work-plan will be prepared in a meeting in June 2004. A first effort estimate will be made available at the next TSB meeting in July 2004.
  • New actions:
    • 28.1 - D.Wright, G.Folger to propose a new Web design and implement a first prototype, where also a more intuitive download page is implemented.
    • 28.2 - J. Apostolakis, G.Folger to put in place an automated system for regularly reminding by email at every quarter to the Geant4 Collaboration about notification of papers published or to be published.
    • 28.3 - G.Cosmo to provide a summary on the benchmark tests performed on the g++ compiler on Linux/Pentium with different optimisation options and chip-specific optimisations.
    • 28.4 - H.Kurasige to provide a summary to Category Coordinators about the foreseen migration of names in G4ParticleChange.
    • 28.5 - T.Sasaki to provide summary of plan for tracking about handling zero kinetic-energy plarticles in along-step at the next TSB in July 2004.
    • 28.6 - J.Apostolakis to prepare planning for Users Workshops to be held in 2004/2005.
New members
  • Approved: Victor Serbo (SLAC, serbo@slac.stanford.edu) in the visualization group to work 50% of his time on AIDA/JAS3 and GAG integration through GUI.
  • Approved: Max Turri (SLAC, turri@slac.stanford.edu) in the visualization group to work 50% of his time on JAS and HepRep2 driver.
  • Approved: Bernardo Tome (Bernardo@lip.pt), Maria Espirito-Santo (catarina@lip.pt), Ana Keating (keating@lip.pt), Patricia Goncalves (Patricia@lip.pt) of LIP (Lisbon) in the EM low-energy group to work 20% of their time on advanced examples, maintenance Bremsstrahlung/photo-electric models, precise angular-distributions and support.

Upcoming releases

Release 6.2 schedule
  • GC presented the schedule for release 6.2, announced already on early March 2004.
  • The Geant4 dev area on AFS at CERN will be reorganised, following new package installation conventions. An announcement with the details will be distributed:
    • gcc-3.2.3 will replace gcc-3.2.0 and will become de default.
      gcc-2.95.2 (the old default) will still be available.
    • The default CLHEP will remain 1.8.1.0. Porting will be done also on the new version 1.9.0.1 as soon as it will be released.
  • The software will be tested also on the new CERN/Linux CEL3 based on Enterprise RedHat Linux, as soon as it will be available
  • No support will be officially guaranteed for RedHat 9. First level testing for releases may be done. FJ reported that TRIUMF is willing to provide a machine which can be used as remote system by STT. SS and FJ will follow up this.
Reports from Working Groups on progress
  • Multi-category:
    • Support for Windows DLLs is still on time for release 6.2. The objective we would like to achieve is to have shared-libs as the default for all platforms in 7.0.
    • Configuration: a switch to -O2 optimisation for Linux-g++ maybe considered for release 6.2. From first benchmarks made by GC, libraries/tests compiled with -O2 on gcc-3.2 resulted in stable results with an average performance boost of ~15%. An alternative G4SYSTEM configuration for Linux (Linux-gO2) will be provided soon for tests on physics.
  • Geometry:
    • Graphical enhancements to the OLAP tool is in danger. The developer responsible has left CERN. Looking for substitute. Item shifted to 7.0.
    • Twisted trapezoid: the framework for twisted shapes will be in release 6.2, together with specific twisted-tubs and eventually box shapes.
    • Field: a new advanced examples testing the field in different configurations will be provided (purging_magnet)
    • The rest of the scheduled items are well in time
    • A new developement not scheduled may be included, as contribution from I.Hrivnacova (IPN/ParisXI): enhancements to the reflection factory and divisions for the reflection of divided volumes.
  • EM standard:
    • In general on schedule.
    • Some problems reported are under investigation.
    • A 10% improvement in performance is expected in some areas.
  • EM low-energy:
    • In general on schedule.
    • New improved model of PIXE is expected in 6.2 instead that 7.0, due to user request.
    • New models for fluctuations (p/ion) is shifted to 7.0 instead.
  • Hadronics:
    • More or less on track.
    • The list of planned activities on the Web for hadronics requires to be updated !
    • Improvements in performance for binary-cascade is expected.
    • A signal-handler was added in release 6.1.
  • Particles:
    • No information. HK not present
  • Tracking:
    • Handling of electric field has been added as milestone for 7.0 Zero kinetic energy should be checked by tracking after the along-step action, then never call post-step with zero energy. In hadronics a protection will be temporarly added.
    • To investigate: do we have any use-case/reason for which the final energy can be zero ? In biasing ?
  • Run, Event & Detector Response:
    • Going according plans.
  • User & Category interfaces:
    • Going according plans
    • Usage of Python to steer Geant4 should be studied. A user proposal posted in the "Geometry" Hypernews Forum will be investigated.

Technical Forum

  • A Technical Forum meeting will be held in Vanderbilt on May 12th:
    • Category Coordinators are invited to review the list of pending items in Makoto's summary presented at the last Technical Forum.
    • Updates to open items should be reported to J.Apostolakis and/or Petteri.
  • HPW makes notice that attaching the Technical Forum to Workshops is a good idea to make it more effective.
  • MGP and JA agree in the idea of organising User Workshops as "Conferences" with timely announcements and promotion. Have a regularly one User Workshop per year. A User Workshop at CERN is encouraged by CMS!
  • Technical Forum meetings may be in future organised differently, for example focussed on some aspects/themes (physics, kernel, software, ...), so that not all (and the same) people/experts have to be present at every meeting.
  • GC makes notice that also a proper education to users must be given when requirements are advanced personally as emails to coordinators. Coordinators should point them to use the existing tools (UR Keeper) and formulate the requirement.
  • Next Technical Forum: may be organised at CERN immediately before CHEP04. Other possibilities: the last afternooon of CHEP04 or within a slot attached to an ATLAS week...

Reports from member experiments/organisations

  • ATLAS:
    • Geant4 is ready since time, date for start of production is delayed of one month
    • Small production runs based on 6.1 made before Easter manifested two major problems (NOT occuring in 6.0.p01):
      1. 20% failure rate for jobs in muon pair-production, mostly non reproducible. Last investigations led to one case reproducible which is currently under study
      2. Particles getting stuck between surfaces with zero step-length (1-2% of the jobs) in a combination of G4Polycone and G4Cons.
        Is it the case to introduce a protection against cases of zero step length (in the order of >100 steps) ? A problem report has also been recently open in Bugzilla (#620) for rare cases (1 out of 10^7 incident electrons).
      Problems are under investigation.
  • LHCb:
    • Happy with 6.1. The lowest failure rate ever.

Conferences and Publications

  • ICHEP 2004 (May 14th) in Bejing. Only collaboration members from the hadronic WG are going.
  • IEEE Conference in Rome in October. There will be also a short 1-day course of Geant4 as for last year in Portland. The fee is rather high. A concern about the fee for the Geant4 course should be written and sent to IEEE. HPW will write a draft, MGP will take care to forward it to the IEEE conference committee.

Workshops and Tutorials

  • Geant4 course in Alghero (Italy) for the National school in June.
  • It would be good to have Geant4 lectures present in the CERN Summer Students programme and/or in the CERN School of Computing (the last time being in Greece in 2000). To be planned for 2005 ?
  • 80 people assisted at the SLAC Geant4 Tutorial in March.
  • The programme for the Geant4 Workshop in Catania needs to be drafted as soon as possible.

Collaboration Agreement

  • Version 7.02 has been circulated to the TSB. Some conflicts with the previous version need to be resolved. A votation on a "key" aspect has started. The objective is to draft the "TSB amended" version 8 by end of May.
  • HPW states that the decision should be taken by "consensus" and not through vote. Seconded by MGP. HPW volunteers to bring people together and look for a concrete proposal by consensus.