Simulation Project Leaders Meeting 14-October-2004 ==== Attendees Paolo Bartalini (PB), John Harvey (JH), John Apostolakis (JA), Witek Pokorski (WP), Alberto Ribon (AR), Gabriele Cosmo (GC). Excused: Alfredo Ferrari --------------------------------------------------------------------- ==== Minutes Alberto Ribon is welcome at the SPL meeting as the new leader for the Simulation Validation Project. Witek Pokorski has been nominated new leader for the Simulation Framework Project, replacing Andrea Dell'Acqua in this role. There will be a comprehensive LHCC review on November 23rd. A time slot of 40 minutes will have to cover the simulation project and will have to include also feedback from the LHC experiments. The report on the simulation project will have to be ready by Obctober 28th and has to be delivered to LHCC by end of the month. The report should include also coverage of the planned milestones and possibly also an outlook for future milestones. GC will contact the project leaders to provide feedback relative to their projects. --------------------------------------------------------------------- ==== Generator services - Paolo The project has been presented CHEP-2004; emphasis was mainly put on the Work-Package 1 (GENSER) and 3 (MCDB). GENSER: - The work for the production quality version has started. Release 1.0.0 is expected by mid-December. - It is agreed that next version of GENSER will support SCL3 Linux. - Concerning Sherpa, Shasha and F.Krauss are working together for the integration in GENSER which is expected in the first quarter of 2005. From mid-October Shasha will be replaced by S.Makarychev. - Work is going on Photos, M.Kirsanov is working on the inclusion of the most recent version. - For Tauola the situation is more complex since there're different versions available and is not yet clear how to handle them. S.Slabotpisky will concentrate on the integration of Tauola in GENSER starting from mid-December. - One of the authors of CASCADE, H.Jung, has been recently contacted. - Discussion with Peter Elmer about EvtGen (the B-decay package) has taken place in Interlaken: Anders Ryd (one of the main authors) has expressed the interest to participate at the development of EvtGen for LHC; this may change the evolution of the package towards the direction of unification of the different versions. It has been proposed also to held the next EvtGen meeting at CERN next year. - Usage of Root in GENSER is limited to some examples. Examples may be placed in a separated parallel repository; this could be considered in the context of a separate project of "validation" for GENSER. If agreed this reorganisation may happen already for GENSER 1.0.0. - A meeting focussed on management of GENSER releases will be hold after this meeting; the objective is to define a relase/tag policy trying to follow a model similar to the one adopted in Geant4, but tailored to the GENSER project. Also roles and responsibilities for the librarian and "liason" persons will be defined. MCDB: - Work has started for the simple level production framework, two persons are working in Santander to interface the library. - There's now a Working server for MCDB. Only Mozilla browser is supported for the moment, it is in plan to extend support to IE and Netscape as well. In the short term the work will be focussed on the development of the 'search' functionality. - The definition of the Meta language XMHEP (XML oriented) is in progress and it is expected to be completed by end of the year. This has to be considered as an effort for the proposal of a new standard. Documentation needs to be completed still. In the long term more emphasis will be put on the development of the APIs for the different simulation frameworks. --------------------------------------------------------------------- ==== Fluka - Alfredo No reports. --------------------------------------------------------------------- ==== Geant4 - John The 9th Geant4 Collaboration workshop was held in Catania (Italy): - Interesting developments were presented by ATLAS and CMS in the context of fast parameterisation using as common base GFLASH; also new requirements to Geant4 have been expressed in this area mainly from the CMS side. - GRID: a web portal for launching Geant4 applications was presented. - Migration plan for release 7.0 of December were discussed: + migration impacting users on user-defined processes. + migration to use in place of . + migration to the new versions of CLHEP. - Concerning CLHEP, version 1.9.X will be supported for the future development tags, pending official release of the CLHEP library; concerns were expressed from developers for the migration to CLHEP 2.0.X where CLHEP namespaces are forced also for the system of units. The impact for this will be evaluated and a final decision if to migrate to 2.0.X for release 7.0 will be taken shortly. - Platforms: it was agreed that support for gcc-2.95.2 and VC++6 will be dropped. - License: there's need to clarify the position of CLHEP concerning the license, since this may affect also the license for Geant4 (LGPL or BSD). Acceptance suite module: - The development of the comparison program is completed. The mechanism for running the suite regularly at every release is likely to require large CPU time; there's need for more CPU resources (GRID?). Upcoming developments: - A new patch to release 6.2 (patch-02) is under preparation. The most important fixes included concern the usage of parameterised materials and a problem in memory management for the Bertini-cascade hadronic module. --------------------------------------------------------------------- ==== Simulation framework - Witek GDML: - The schema has been extended and now it's almost complete; support for replicas and parameterised volumes has been added (tested on Geant4 novice examples N02 and N04); still missing divisions; added polycone, to be added polyhedra. - The reader/writer module is now complete. It has been tested on a real detector geometry, the LHCb detector can be loaded and exported: GDML ASCII file of the order of 2.8 Mbytes (~5000 volumes); good performance in loading in memory. - There's big interest on GDML from external Institues (Fermilab, ESA, Los-Alamos, ...) which gives good motivation to bring foward the project. Modularization of the geometry description and the strong interest on having a human-readable language decoupled from the C++ programming language are the strong arguments in favour of adopting GDML. - Next steps are: + complete the schema to support the missing shapes and test the import/export feature with a more complex geometry like the ATLAS detector; + implement plugin for Root. Python: - Need to understand how to "pythonise" Geant4. Strong concern was expressed at the Geant4 Workshop about the need of protecting public interfaces in Geant4 first to consider this use-case. FLUGG: No much news on this area, most of the activity in the past month was focussed on GDML. WP had a discussion with Paola Sala about how to 'mimic' sensitive detectors in FLUKA for the purpose of providing a working example in FLUGG. No much interest has been expressed from experiments in using FLUGG; ATLAS through Fabiola has expressed interest in having validation activity with G4 and FLUKA through FLUGG for the calorimeter test-beams. --------------------------------------------------------------------- ==== Simulation physics validation - Alberto The first phase of the hadronic validation has been concluded (10% agreement between Geant4 and data). For the simple benchmarks, the validation has been carried out for both Geant4 and FLUKA against experimental data. The note on PI-absorption is ready; waiting from feedback from the Geant4 experts before publishing. Concerning the calorimeter test-beam, the comparison has only been made for Geant4, it is planned to use FLUGG for extending the comparison with FLUKA. Future plans: - Electromagnetic physics: perform vlidation at the permille level, following interest expressed by ATLAS. - Hadronic physics: definition of a new simple benchmark to help also for a better understanding of the shower shapes. Two solutions are possible in this respect for improving the comparison between simulation and data: either perform a tuning of the xsections in the simulation (between the experimental error bars) or wait for more precise and better understood test-beam data and perform again the study. - Although at the moment no interest in heavy-ion simulation validation was expressed, this may become a future activity. Man power issues: - Manuel Gallas is involved in ATLAS test-beam data and validation; first preliminary results are expected for beginning next year. - For the simple benchmarks the only persons involved are currently AR and WP. It would be desirable if experiments could help in this area. --------------------------------------------------------------------- ==== AOB JH reports that concern has been expressed about future support of Garfield (Fortran package for detector response mainly in gaseous detectors). Issues of long term support and maintenance need to be clarified, as well as for potential re-engeneering of code. There may be the possibility to have Rob Veenhof as scientific associate to work on it. Next meeting to be scheduled by mid-November.