Minutes of the Geant4 Steering Board meeting 62
9th December 2010
**Attending **: Makoto, Koichi, Gabriele, Marc, Peter, Joseph, Hisaya, Takashi, Dennis, John, Gunter, Vladimir, Daniel, Paul.
Excused: Christina, Sebastien.
1. Regular Items :
a. Approval of the minutes of SB meeting of October 8th, 2010:
b. Updated action items as of October 2010.
2. Report from Technical Forum on November 16th 2010:
The TF agenda can be found at : http://indico.cern.ch/conferenceDisplay.py?confId=112827
a. CMS report:
CMS production will rely on Geant4.9.3(4) (+ patches).
CMS targets to have homogeneous Monte Carlo sample. They plan to change Geant4 version for:
– Bug fixes which do not seriously impact physics results
– Performance improvements
The December release 9.4 is targeted as to be the one for massive production in 2011. CMS has plans to migrate to 9.4 (+patches). They will use “whatever will be available to make a stable production cycle. The migration will depend on the results of their validations. At present, validation is ongoing with the latest reference tags and they do not foresee major issues.
Request to Geant4 for a major release one month before the Christmas break:
During the December 2011 & January 2012 shutdown, the experiments get the opportunity to change their detector and software running conditions. They prepare also for large MC (and data) production with the latest version of software, with a production starting from late December / early January. This schedule leaves very little time to validate the latest version of Geant4. Hence the request for a Geant4 schedule release one month earlier.
This request is seconded by LHCb and ATLAS.
It is proposed to have the final candidate tag by mid-November so that LHC experiments can already start their validation. It seems such an early candidate used to be proposed in the past. It is also pointed that this may be a benefit in helping figuring out issues earlier. An anticipated date for the 2011 December release would also fit well with the possibility to have the next Geant4 Collaboration Workshop held earlier at SLAC for fiscal year restrictions.
Gabriele Cosmo will provide a calendar to meet a first of December 2011 release date to evaluate the request.
b. LHCb report:
LHCb used release 9.2 in production during 2010.
LHCb preliminary reports of 15-30% difference in EM energy scale between 9.1 and 9.2, with of 9.2 results similar to that of 7.1.
News since TF meeting:
- Plan to continue 9.2 in production during 2011.
- No update received from LHCb on EM issue scale between 9.1 and 9.2. It is suspected that LHCb is comparing different physics list setups.
Vladimir Ivantchenko to follow the 9.1-9.2 LHCb energy scale difference issue and check for improper physics list(s) usage or real physics issue.
c. ATLAS, ALICE reports:
No relevant issues reported and discussed.
LHC experiments agreed to the proposed end of support for Geant4 9.2 in December 2010. All experiments have planned to migrate to release 9.4 in early 2011, this will be ensured in view of guaranteeing proper support.
3. 9.4 release:
Gabriele reports that the last candidate tag for the release is under preparation. Developments are all included.
Open issue: performance hit of 3%, apparently due to cross-sections. If a simple fix is available, it could be incorporated
Andrea Dotti will report on the last performance measurement on release 9.4, including findings on the potential 3% CPU penalty.
4. Migration to SVN:
a. Proposed schedule (Gunter):
Gunter reminds that a constraint is that Przemyslaw Adam Paprocki, student who provides the technical support, is available only until February.
Gunter proposes to move the repository on the week of January 10th; he anticipates that at least 24 hours will be needed.
Close CVS repository in the afternoon of Monday, January 10th. Benchmark and verification on a new g4test repository from Friday 10th of December up to Monday 13th.
Migrate documentation after source code.
Migrate Geant4 Tools (small) and LXR in week 13th – 17th of December.
Gunter reports that the new tool to replace Bonsai tag system is now available. Several developers have already proposed tags through this system.
Gunter will distribute a guide for most useful SVN commands, he informs that the page is ready, and will be distributed after the meeting.
b. Agreements to be “signed” (Makoto/Marc)
Makoto and Marc will provide a text by end of this year for agreements to be signed for SVN lock.
It is proposed to use this lock system for membership control and to keep track of who is active by signing on first access of the year to SVN - agreed.
5. Architecture revision:
a. Final review document (John):
The final review document has been provided by J.Apostolakis provided. Only cosmetic changes have been made since October. There was no enough time to revise more.
It is proposed that comments are to be provided by Wednesday time (15th of December) – agreed.
Dennis proposed the new prototype system.
If no objections are raised in the next 24h, the proposal will be approved. Dennis underlines the change from 2 to 3 weeks for internal submission deadline.
7. Improving credibility / collaboration-wide effort:
a. Requirements tracking system (Koichi):
The JIRA system has been tried. A server is under preparation, and might be ready by January. A report is expected by next SB.
b. Unattended Bugzilla reports (Koichi):
It is proposed that if a bug is assigned to a particular person who is not a WG coordinator, the report should be sent to this person, and cced to WG coordinator.
c. HyperNews postings (Makoto):
Can affiliation of sender be obtained when responding to message? At this point, need to ask users to put in their affiliation. No conclusion.
d. Sharing system testing (Gunter/Makoto):
8. Preparation of 2011 Collaboration processes:
The Geant4 census target date is 18th of February (Makoto).
Makoto will write to WG coordinators and mention about the SVN lock for membership renewed once per year.
The target date is accepted, and acceptation to proposal for membership.
a. New collaborators:
Aida Galoyan : membership accepted. She will be working in the hadronics WG on anti-proton interactions with protons and nuclei.
b. Near future conferences and events:
Geant4 Course Tutorial: Oak Ridge National Laboratory/RSICC, Oak Ridge, Tennessee (USA), 7-11 March 2011.
First Geant4 Australian School and User Workshop: University of Wollongong, Wollongong (Australia), 11-16 April 2011.
c. 2011 Collaboration workshop:
SLAC is still working for fixing date/location.
Thursday, December 9, 2010 - 12:00