Geant4 TSB Meeting

Submitted by Anonymous (not verified) on

19th June 2003 9:00 AM 4/S-013


Meeting called by: John Apostolakis

Note taker: John Harvey

Attendees:

@Annecy : M.Maire

@CERN : M.G.Pia,S.Sadilov, A. Dell' Acqua, A. de Roeck, V.Ivantchenko,

J.Apostolakis,J.Harvey,G.Cosmo, P.Arce, G.Folger

@KEK : S.Tanaka

@Manchester : J.Allison

@SLAC : D.Wright,M.Asai

@Stockholm : Irena Gudowska

@Triumf : F.Jones

Extended Minutes

Draft 1, 25th June 2003,

1. Minutes and actions from meeting held on March 18th 2003

The minutes of March meeting were circulated just before the meeting. Please send comments to John Apostolakis within the next week, to finalise these by end-June.

John informed the TSB that he has received the resignation from Youhei Morita, who was convener of the Persistency category and has now taken up new duties at KEK. A replacement will be sought.

Action items from the March 18th 2003 meeting:

1.[WG coordinators] - Publish the WG work-plan for 2003- this is largely complete and status can be found on the web. For some categories the plan is not yet complete e.g. persistency and visualization. In a number of cases there is a need to review the resources available in concert with this planning. In the missing persistency plan it was proposed to add an implementation using POOL and resources must be envisaged for this.

2. [G.Cosmo] - Update release procedures to document that minor releases *may* include changes of interfaces internal to Geant4 -- requiring that users who have created their own processes must review and potentially modify their implementations. (Appropriate documentation must be made available to assist these actions.)Release notes of Geant4 5.2 included explanation, but procedures not yet revised. Still open.

3. [ John Apostolakis ] : Create a project to review the performance of Geant4 for large detector simulation, including contact and technical persons from experiments with current requirements on performance and the relevant Geant4 experts. Actions under way to review performance - a project is not set up. Still open

4. [ John Apostolakis/Katsuya Amako/CA-editors ] Circulate revised proposal for Collaboration agreement to TSB and CB by April 15th. Draft circulated. Undertaken. Closed

5. [ John Apostolakis ] : Form a delegation with a mandate to represent TSB / developers at meeting with CB representatives regarding the Geant4 Collaboration Agreement. Undertaken. Closed

6. [ Gunter Folger ] : Perform a 'full' survey of platforms and compilers in use mid-2003 and projected for end-2003 in partners. Unclear overlap with previous survey. Still open.

7. [ Representatives of laboratories/experiments/institutions ] : Provide information on currently used platforms, and those planned for end-2003. Note that a partners requiring support for 'fringe'/unpopular platforms will be expected to contribute to this support (or undertake it as a contribution, if they are the sole requester.) As action 6 above. Still open.

2. Collaboration Agreement

Meeting of TSB developer representatives and CB delegates formed an adhoc committee, which met May14-16 regarding the CA. A revised proposal was drafted and sent, first to the TSB and then to the entire collaboration for comment. John reported that he has received a small number of comments so far, generally positive. HPW had circulated his comments before the meeting, stating that membership of the Steering Board is so far ill-defined, and that work is still needed to define a publication policy. Florence expressed the wish that we do not enter a new discussion on the proposed CA. John reminded the meeting that, in their original mail, John and Richard had requested that only the minimum, absolutely necessary, changes be considered. John H. asked whether a timetable had been set to produce a new (hopefully final) revision.

New Action item : John will communicate with Richard Mount to find out the current status of the Collaboration Agreement document and will circulate new information to the TSB

In response to a question from Albert, John stated that it could take until the end of 2003 before the new Agreement was actually signed and the new bodies could be brought into action before this. In order to keep momentum it was agreed that pre-meetings of the new bodies take place in the interim period.

New Action item : John to find someone to organise a pre-meeting of the new Technical Forum within the next few months. Vancouver represents a natural opportunity. It was agreed that a Technical Forum at CERN soon after is also needed.

New Action Item : John to follow up HPW's concerns on publication policy.

In response to a question from Pedro, John underlined that it was intended that collaboration members (and my extension those co-opted to the Steering Board) would be people who contributing to the development, maintenance and testing of the toolkit, and not users of Geant 4. The Technical Forum is the body intended for user representatives.

New Action Item : The TSB mandated John to inform the CB that the TSB looks on the new Collaboration Agreement proposal favourably.

3. Geant 4 Workshop

Fred informed the TSB that arrangements have been circulated to the collaboration in a first mail shot. A website has been setup and comments are welcome:

Website : http://www.triumf.ca/geant4-03/and emailgeant403@triumf.ca

Peter Gumplinger is absent till July 7th. In the past a small number of advanced users had been invited to the collaboration workshop to promote technical discussions. Please bring proposals to Fred (for American continent) and to John.

New Action Item : Fred and John to solicit advanced users to participate in Workshop.

The programme for the workshop has not yet been defined. A discussion took place on possible themes. Topics proposed included :

- validation of Geant 4 physics, comparison of Geant 4 results with testbeam results etc.[John]

- usability of Geant 4, documentation, support, training, interactivity of tools[Maria Grazia]

- a project to disseminate the use of the tools for statistical testing and to use them to revise a small number of physics tests and test beams that comparing Geant 4 results with physics results.[John]

Maria Grazia pointed out that the framework required for physics testing is different from that of system testing. This could be an item for further work or a project to be discussed in Vancouver.

4. Progress with Release and 2003 Objectives

Gabriele reported on the current status of the 5.2 release (scheduled for end of June). The testing of all candidate tags submitted in the nominal period for groups of categories has been accomplished. In addition the major migration to direct use of STL, in place of wrappers that work on the non-standard STL of the egcs compiler, has been achieved. A small number of tags are still waiting to undergo system testing. Memory leaks observed in hadronic models and remain to be fixed. Other bugs that caused crashes in at least 1 physics list have now been fixed.

Maria Grazia referred to the release of a new physics process that caused problems to some users, in particular GLAST. This highlighted the need to improve the testing and verification procedures before deciding what goes into a new release. Discussion between category coordinators is also needed before a new release is made. The experiments also provide important test platforms. Albert, Andrea and Florence underlined the importance of good quality release notes in order to understand the changes that have been made so that appropriate testing can be organized by users.

New Action Item : John to organize a discussion between category coordinators at the next TSB in Vancouver and to come out with a proposal on how to improve the release process

New Action Item : Gabriele to give a presentation at the Workshop on the current process.

5. Reports from WGs on progress towards 6.0 release

A full discussion on this topic will take place in Vancouver. Gabriele stated that the first development tag after the end June release will be end July.

Makoto pointed out that advanced users may need to make changes to accommodate the new Run Manager, ordinary users wont. Full documentation and examples will be provided.

Maria Grazia said that a first release of the new statistical testing tools will be made end of June. There then followed a discussion on the right approach to testing. Gabriele said that the first stage is to integrate the test suites in unit tests, and only then consider their use in system tests. Gunter suggested that a presentation on testing procedures is made at the Workshop. John added that we should also have a talk giving the first experience of users who have tried them (2-3 beta testers).

New Action Item : John to put "Use of Statistical Testing Tools" on the agenda of Workshop and to identify beta testers and someone from system testers to present their experience.

6. Reports from Member Experiments

Florence

 reminded the TSB that LHCb would like to use Geant 4 on Windows for development, which requires the use of shared libraries. A set of identified proposed changes would need to be undertaken to make it possible to use sharable libraries on Visual C++ 6 and .Net . Gabriele reported that that an initial problem (due to the use of static in-line code) had been fixed, but that specific Windows treatment in a significant number of sections of the code would be required to address these issues.

CMS Albert asked if attention could be given to memory usage. Currently OSCAR, the CMS Geant4 simulation, appears to use between 200-500 MB needed for typical runs. Initial measurements attribute about 80 MB to physics lists with Geant4 5.1, when utilising a large number of materials. An effort to revisit the materials budget was under consideration within CMS, and assistance from Geant4 in configuring and minimizing this was requested.

It was pointed out that the energy range and number of bins for tables of physics processes are already dynamically configurable (at runtime) and this can help to minimize memory usage.

Imperial College London would like help to understand the impact of the energy cut-off on delta ray production (currently 250 eV but would like 100 eV). Vladimir stated that reducing the cut-off represents a major new development i.e. a new project which would need manpower to complete. Nevertheless IC would like guidance in interpreting the significance of the results they obtain with the current cut-off. They will contact Vladimir off-line to understand this better.

7. Geant 4 Publications

John Allison was warmly thanked for shepherding the Geant 4 paper into publication. John will provide 1 bottle (magnum?) of champagne (brandy?).

Maria Grazia summarized her proposal to launch a project to create a Geant 4 Physics Book. The purpose is to make a standard reference on Geant 4 physics and its validation in an experiment context. Such a document would be invaluable for developers, students, and even funding agencies, as well as acting as a focal point for many future activities and helping to engender team spirit. It would be a huge effort but rewarding. As a point of reference, the BaBar physics Book had ~70 contributors and took ~ 1.5 year to produce. A first discussion has been organised for June 26th at CERN to take advantage of a visit by Katsuya Amako. John recommended that a feasibility study be made, to understand issues of timing and resources. HPW had circulated a comment before the meeting that he supported the concept but that the timing is not right for hadronic physics. Michel agreed that such an endeavour should be made if the effort can be found.

New Action Item : Maria Grazia should report to the next TSB the outcome of the meeting on Geant 4 Physics Book to be held on June 26th.

8. Geant 4 License

John reported that the signing of a new Collaboration Agreement was a great opportunity to resolve the issue of the Geant 4 Software License. In order to prepare the work needed to produce a license John asked that two pieces of information be collected.

New Action Item : John to solicit information from Institute representatives and independent contributors on their policies regarding who holds copyright on the contribution of their members. Also to solicit the input of working group coordinators on the list of contributors to their area of Geant 4 and the approximate extent and longevity of their contribution.

9. AOB

It was proposed to make the Delta Review Report more visible by having a link to it on the Geant 4 main web page.

It was reported that a Geant 4 training course will be held in Portland in October

Summary of Full List of Action Items Remaining After the Meeting

1.[WG coordinators] - Publish the WG work-plan for 2003- supply missing pieces such as visualization and persistency.

2. [G.Cosmo] - Update release procedures to document that minor releases *may* include changes of interfaces internal to Geant4 -- requiring that users who have created their own processes must review and potentially modify their implementations. (Appropriate documentation must be made available to assist these actions.)

3. [John Apostolakis ] : Create a project to review the performance of Geant4 for large detector simulation, including contact and technical persons from experiments with current requirements on performance and the relevant Geant4 experts.

4. [ Gunter Folger ] : Perform a 'full' survey of platforms and compilers in use mid-2003 and projected for end-2003 in partners.

5. [ Representatives of laboratories/experiments/institutions ] : Provide information on currently used platforms, and those planned for end-2003. Note that a partners requiring support for 'fringe'/unpopular platforms will be expected to contribute to this support (or undertake it as a contribution, if they are the sole requester.)

6. [John Apostolakis ] Communicate with Richard Mount to find out the current status of the Collaboration Agreement document and circulate new information to TSB

7. [John Apostolakis ] Find someone to organise a pre-meeting of the new Technical Forum within the next few months. Vancouver represents a natural opportunity. It was agreed that a Technical Forum at CERN soon after is also needed.

8. [John Apostolakis ] Follow up on publication policy.

9. [ John Apostolakis ] Inform the CB that the TSB looks on the new Collaboration Agreement favourably

10. [Fred Jones / John Apostolakis ] Solicit advanced users to participate in Geant 4 Workshop

11. [John Apostolakis ] : Organize a discussion between category coordinators at the next TSB in Vancouver and come out with a proposal on how to improve the release process

12. [Gabriele Cosmo] : Give a presentation at the Workshop on the current release process.

13. [John Apostolakis ] : Put ?Use of Testing Tools? on the agenda of Workshop and identify beta testers and someone from system testers to present their experience.

14. {Maria GraziaPia] : Report to the next TSB the outcome of the meeting on Geant 4 Physics Book to be held on June 26th.

15. [John Apostolakis ] : Solicit information from Institute representatives and independent contributors on their policies regarding who holds copyright on the contribution of their members. Also to solicit the input of working group coordinators on the list of contributors to their area of Geant 4 and the approximate extent and longevity of their contribution.