Notes of 2nd Geant4 SB meeting
Meeting Minutes
6 February 2007
Document Version 1.1.6, 14 Feb 2007, 16:10
CET
Opening:
The regular
meeting of the Geant4 Steering Board was held,
starting at 8:00 CET on 6th February
2007 by VRVS videoconference.
SLAC: Dennis W., Makoto A., Joseph P.
CERN: John Ap.(chair), Gunter F., Vladimir Iv., Maria-Grazia P., Gabriele C., Witek P., Ian McL.
KEK: Takashi S. (1st half), Koichi M.
Excuses: Katsuya A., Hajime Y.
The
minutes of the October 13th meeting were approved.
Updated separately (Gabriele C.)
Input on issues, new and existing, from Geant4 Technical Forum of 31st January 2007:
-
recent reports
on significant failure rates running G4 8.1 (CMS) and
8.2 (ATLAS)
from experiments; likely cause of the problem has been identified
,
involving the interaction of multiple scattering and field propagation in the
presence of overlaps in the model geometry.
An assessment of fixes already discovered and further investigation are underway. Additional fixes and improvements are expected. Any remaining issues to be addressed at a technical meeting February 20th or 27th.
Identified the need to extend testing to cover use cases with more complex geometry setups, to test geometry and physics robustness.
Extend category and system integration tests to include some with a number of overlaps in the geometry in order to test the behavior of the toolkit Encourage experiments to do regular and routine testing of their geometry, to identify significant overlap problems – before utilizing a geometry in production.
Issue: Are experiments utilising the development releases (eg provided on AFS at CERN) to undertake porting and testing of new developments and testing ?
1. (Vladimir Ivantchenko/Gabriele Cosmo) Report on progress and resolution of problems causing the aborting of events in large setups when using Geant4 8.x and new multiple scattering (e.g. in QGSP).
2. (Witek Pokorski): Obtain a full (large) HEP experiment setup, using a realistic geometry taken from a recent LHC detector description (including potential overlaps) and create a simple corresponding test.
Note: many items were circulated or posted only in the 24 hours before the meeting, making it hard to assess them carefully in time for the meeting.
Vladimir Ivantchenko expressed concern about use of some physics process in cross section biasing. Potential for incorrect results when using particular processes.
New action:
Vladimir Ivantchenko to identify the issue and the relevant people.
Developments for the tunable tolerance ready, will be
included in the development release
New action:
4. Gabriele will notify category coordinators about the revisions involved in enabling tunable tolerances, which allows changing the value of kCarTolerance in a Geant4 job
Generic Processes & Materials: question to Marc (missed - TO DO)
Question arose regarding of the implementations of Bremstralung and polarized photoelectric processes. It was agreed to discuss these separately, and planned developments between the EM groups, to foresee potential commonalities or new requirements.
The naming of EM components of hadronics physics lists (PLs) are currently confusing, and will be revised.
The developments of the (separate) Geant4-DNA
project, will be provided inside
the
low-energy EM package in the toolkit.
Concern was expressed regarding the proposed implementation
of parallelization, using MPI. Proposed to hold a
discussion on this, to seek commonalities .
Foresee a user inquiry about and a review
of platform, for use and testing, for the next TF meeting.
Documentation
management: circulated alpha-version of DocBook documentation conversion. Some
feedback was given.
The proposal for a reverse
Monte-Carlo integrated in Geant4 was briefly summarized. Potential common elements with Geant4e are being sought, and a
technical discussion is being arranged.
A revised format for the workshop was
agreed, with the first three days,
Thursday-Saturday
open to users and days 4-6 (Mon-Wed) a Collaboration Meeting, exclusively for
collaboration members
Day 1 is intended
for user presentations, and Geant4 presentations on current
status and developments.
Days 2-3 are earmarked for
matters
which can involve or are particularly relevant to users. .
A Program
Committee (PC) will be set up. The PC will oversee the workshop
program. For aspects of particular local interest, this will be undertaken
together with the local organisers (LOC). The SB chair is to
nominate the members, and the SB to approve.
Volunteers are requested.
A small executive International Organizing
Committee (IOC), including the chair of the Local Organizing
Committee (LOC), will oversee other organization matters.
The Steering Board chair will appoint
.
- A key
initial consideration for the Program Committee's is to determine whether
tutorials should be offered.
The following
are in the form of a proposal, about which comments should be received by 16th
Feb.
- The workshop
venue will be available to user groups on the Wednesday preceding the workshop.
Users would pay for the hire of rooms on the Wednesday.
- The local scientific
societies will be invited to particular open
sessions, and welcomed in user-related sessions, as agreed,by
the LOC and IOC.
- The fee will cover
lunch and refreshments, administration, and an
optional social event. Two options will be offered: one for users
attending the first three days, and another for collaborators attending the
full meeting.
The review committee has been constituted and the
mandate is being finalized. Reviewers will include experts covering HEP,
medical and space applications. A web page is in preparation, and will be circulated
to the SB.
Practice talks
will be arranged to provide feedback already from the collaboration.
Proposed schedule
is
during the last 10 days of March
The reviewers will contact the collaboration
concerning specific issues.
We need to prepare also material
addressing areas in the mandate, including validation,
physics performance, time performance, and usability.
A proposal has been received from a Fermilab team to join the collaboration,
proposing to work initially on hadronic physics validation,
numerical stability and time performance.
Some members expressed interest in
seeing new teams assume
responsibility
for
part of the development as well.
The vital need to have excellent
communication with physics model
developers, in order to contribute to improving
the models was also noted. When potential discrepancies arise during validation,
major
effort is involved in finding the root
cause(s).
The SB agreed to adopt the process
proposed by J. Apostolakis for the SB’s examination of the Fermilab proposal. J.Ap.
will proceed to implement it, appointing two G4/SB members to assess the
proposal in collaboration with the WG coordinators in the relevant areas, the
SB chair and deputy.
Maria-Grazia requested
restricting access for a new collaborator
to parts of the CVS repository outside his/her working group(s)’s
responsibilities.
The proposed process is agreed.
The SB agreed to launch it immediately after the
review (end of April).
The selection of the nominating committee can now commence.
AOB:
Action
on Aatos to collect comments on IAEA initiative.
Adjournment:
Meeting was adjourned
at 12:10 by J Apostolakis. The next SB meeting is CET
on 13 March 2007 in. An initial reservation for 15:30 CET has been made in
Initial version of minutes
by: Witold
Pokorski, 6 Feb 2007
Revised
by:
John Apostolakis, 6-14 Feb 2007