Simulation project leaders meeting Jan 14, 2004 =============================================== === Attendees Alfredo, Fabiola, Paolo, Torre, John, Andrea === Agenda Around the table on project status and plans for 2004. === Minutes == Physics validation In the process of finalizing the collection of physics requirements. A note will be produced by the end of the month and comments on it requested. It will be a dynamic and regularly updated document as requirements continue to be refined. In simple benchmarks, Juerg finished his study and a note is in limited circulation for comments. It will be circulated to Alfredo, Federico, Hans-Peter and JohnA for comments before a draft is made public. Efforts to find people in the experiments to work on simple benchmarks continue to be unsuccessful, so given the priority of the work, Witek Pokorski (who replaces Juerg from Jan 1) will work on simple benchmarks with part of his time. He has started on the second benchmark, looking at pion absorption. As reported earlier the first cycle of EM physics validation is complete, with the effort now concentrating on specific detailed issues, e.g. recent work has concentrated on shower shapes and a case observed in ATLAS HEC where data-G4 agreement is not as good as it should be. Hadronic physics validation is ongoing using the improved HAD physics lists. A small working group was set up to compare ATLAS/HEC, ATLAS/Tilecal and CMS/HCAL to QGSP and LHEP lists, producing the same plots for all subdetectors, to draw some conclusions on which physics lists are good for what. A report summarizing the conclusions will be made at the Feb 4 physics validation meeting. This is the first example of the planned comparisons "across experiments". Restarting the use of FLUKA in the planned tilecal test beam study has still not started but should begin soon. It has awaited the infrastructure to use FLUGG with Geant4 test beam geometries that Andrea and Witek are working on (see below). == FLUKA Alfredo gave a comprehensive report on FLUKA status and plans based on the INFN planning that has recently been done. Many of the goals are specifically directed at CERN needs. An important recent development is that the INFN/CERN agreement was finalized (signed by CERN DG and INFN Council) last month, establishing an official 3-year joint project commencing Jan 2004, with automatic renewal every 3 years if neither party withdraws. The agreement includes a public release of FLUKA code, probably after summer 2004 if no major problems appear. Other 2004 milestones: - Final version of manual. Draft of ~360p exists. Purely technical: how to use the code. Couple months to finalize. Will be one of two reports, other one on the physics, will take more time, should be out this year though. - Courses. Another to come, probably in Houston in the fall. Reactions have been positive. Examples will be collected from the course and put on the web. - Present preprocessor for em physics will be eliminated. At the same time, updating of photon and electron cross-sections is being done, together with a general cleanup. The CERN EP scientific associate is working on this. Should be ready for May release. - Most recent release (just before vacation) supports a new version of FLUKA geometry that is more friendly for users. Geometrical shapes and regions can use names and not just numbers. Two levels of parentheses are now allowed in boolean geometry definitions. A natural extension will be supporting names everywhere in FLUKA input cards; this is an objective for May release. Support for the old way will be preserved. - physics improvements will include further improvement/expansion of evaporation/fragmentation/fission codes to manage production of relatively rare fragments; induced radioactivity prediction; heavy ion fragmentation; and a library of radioactive decays for all possible nuclei which will allow the user to selectively monitor radionuclei produced in a cascade and measure induced radioactivity levels at times of interest; new models for heavy ion interactions below 100MeV/n; and improved integration and ease of use of cosmic ray models which provide cosmic fluxes for the past 40 years and future estimations based on solar cycle and earth position. FLUGG is stable and is being maintained by Paola. The cause of the recently reported ALICE VMC pathologies was found. Andrea asks is Paola will be available to help with any questions or problems encountered in getting the tilecal study going and Alfredo says yes. Andrea asks if he can arrange for FLUKA libraries to be installed in a public area at CERN; Alfredo doesn't see a problem with this and agrees that CERN needs a local mirror. == Generic simulation framework Andrea is actively working on the initial objective of providing a FLUGG based infrastructure to the physics validation project needed to commence FLUKA validation studies (starting with ATLAS tilecal). Witek is helping, splitting his time with physics validation. They are using SEAL components and appwork and should in the coming days offer the first version of an infrastructure that makes both G4 and FLUKA available. The tilecal work should thus be able to start very soon. In the future we will want to extend beyond the test beam physics validation work to using FLUKA in full detector simulation. Will this be done with fully detailed or simplified geometry? If simplified, FLUGG should be OK. For fully detailed we would have to do a validation that the geometry is being dealt with properly, and at that point it would be worth looking at using FLUKA via the ALICE VMC. For that we will need the GDML geometry exchange mechanism to transfer detailed detector geometries to the VMC. We need to start readying GDML for this if this is what we plan. We need to put a workplan for this before the PEB (in view of the reorganization; not the SC2 as previously planned). == Generator services Paolo has already written up project status for the quarterly report so I don't duplicate it here. Paolo already has a detailed plan for at least the first half of 2004 prepared before the last quarterly report. == Geant4 Hadronic physics lists are in the 6.0 release, as is a first version of the abstract navigator. Substantial feedback was received from CMS, ATLAS and LHCb in the 6.0 release phase and this was much appreciated by the release team. The feedback helped to identify issues early, allowing many to be resolved. Some, though, remain under investigation. John notes that rare problems (e.g. only appearing after days of production processing) require special strategies to deal with them. Also in 6.0: volume divisions, visualization of field lines, bonsai improvements (better browser support), a new model approach for em processes (motivated by maintainability and extensibility, and providing a reduced memory footprint), hadronic physics list revisions, new biasing capabilities, and improvements in hadronic modeling, including in cascade models and CHIPS. G4 issues for 2004: - savannah prototype replacement for bug system planned for mid-Feb - new id assignment to particles after inelastic collisions, needed for reconstruction - problem tracing, particularly for rare problems appearing deep into long jobs - wider usage of performance improvement options - a replacement for Steve O'Neale as testing backup person is needed === Planned applications area talks over the next couple months: - Generator services status and plans - Geant4 release 6.0 - EM physics validation - physics requirements review - FLUKA status and plans - Bonsai usage in G4 and possible SPI application - generic framework status and plans