Feb 24, 2004 Simulation project leaders meeting minutes ======================================================= ==== Attendees Paolo, Fabiola, Witek, John, Torre ==== Agenda Round the table on project status ==== Minutes == Generator services - Paolo Bartalini Milestones have been recently reviewed. The next generator services meeting is on Thursday and will feature brainstorming, requirements and design discussions for a simple generator production framework. There will be a review of the generator services project and GENSER in particular on March 25. Michaelangelo Mangano is organizing the review. The review coincides with the regular generator services meeting but with an early (14:00) start time. Current GENSER librarian is Alexander Sherstnev. Recent suggestions from Florence on repository reorganization were appreciated. Currently work is underway on a couple of HERWIG problems with Peter Richardson assisting. Yesterday Paolo and Torre discussed the generator services manpower situation with Les Robertson. This is an urgent problem as the experiments start to rely on products of the project such as GENSER. Lack of stable core manpower also affects our ability to attract and take advantage of volunteers from the experiments because some core expertise is needed to provide support and supervision for volunteers. Les is looking into the problem (which was raised also by the LHCC review). We agreed the milestone for the first C++ generator integrated in GENSER would be suspended until a date can realistically be defined. The plans to integrate Sherpa early ran into problems due mainly to continuity problems in the librarian role. === Simulation physics validation - Fabiola Gianotti Juerg's note is essentially complete and in the stage of collecting comments. Witek is working on the single pion absorption benchmark. This is the second of the two top priority simple benchmarks identified early in the project. Once these are addressed (ie soon) we will review and prioritize the remaining list. John suggests that in assembling the list we should give first consideration to which benchmarks are the most useful to LHC simulation physics validation; other considerations such as data availability and manpower can then be folded in when it comes to prioritizing which should be pursued. The project emphasises again that participants from the experiments are sought to participate in this important and interesting work. The simulation physics requirements review document is essentially ready. The draft will be circulated to the full simulation list today or tomorrow, completing this milestone. Input was received from all experiments. The review will be a working document that will be steadily improved over time. A note drawing first conclusions on hadronic physics validation and addressing in particular the performance of the optimized QGSP and LHEP Geant4 physics lists is in preparation. It incorporates results from ATLAS/HEC, ATLAS/Tilecal and CMS/HCAL calorimeters, with the same plots across the detectors used to compare e/pi ratio, energy resolution, shower longitudinal profiles (where room for improvement has been found), etc. LHCb results are still to come; there is some hope of including them. This note will conclude the first cycle of hadronic physics validation (with the same caveat that it will be an evolving document). Progress is being made on comparing Geant4 and FLUKA (via FLUGG) hadronic performance using ATLAS pixel test beam data. Alberto Ribon is working on this. The G4 part of the study is completed. Hit production using FLUKA is done, digi production from FLUKA hits is in progress. In EM physics, work continues on specific areas in need of improvement. One is EM energy resolution in the ATLAS hadronic endcap; improvements in the latest Geant4 release should improve the situation and a new energy resolution plot with G4 V6 is being worked on by the HEC people. LHCb has observed problems in the simulation of energy response in their Si; they want to switch off delta ray production and introduce a parameterisation which is tricky in G4 so they are getting help from the G4 experts. Fabiola is interested in pursuing with ALICE whether they can profit from the ATLAS work on transition radiation modelling since they also have a TRD (also for showering into the muon system -- profit from ATLAS test beam studies?). LHCb has some first em (Shashlik) calorimeter results with Geant4. EM shower profile results will be collected into a note once the current hadronic validation note is completed. === Generic simulation framework - Witek Pokorski Witek is working primarily on the single pion absorption benchmark mentioned, comparing G4 with FLUKA. He and Andrea (not present) have looked a bit at FLUKA/FLUGG in the generic simulation context but not a lot has been done. Witek produced a draft of a GDML proposal that will now be iterated on. === Geant4 - John Apostolakis One milestone was moved from last year, the initiation of a G4-SPI infrastructure pilot, and this is now underway with a Geant4 prototype of the Savannah portal in place. G4 has identified two key Savannah issues that will need to be addressed before Savannah could be adopted by Geant4. One is anonymous bug reports: the current possibility to submit a bug report anonymously (generally accidentally) should be eliminated. Also, reminder email to owners of long-open bugs should be supported. Both of these have been heard from other Savannah users and are addressable. G4 will prepare a report in the next few weeks with a first attempt at cost/ benefit analysis of adopting Savannah. The 2004 workplans for both the LCG G4 team and the full G4 collaboration are in progress. The technical forum two weeks ago at CERN provided some new input on requirements and prioritization. The principal driver for 2004 plan priorities is the use of Geant4 in production by a number of experiments, with the hottest issues being robustness -- identifying and promptly fixing problems -- and performance. Work items for 2004 will mainly be continuing things with emphasis adjusted for these priorities, without taking away from validation activities. The Geant4 physics book proposal that has been discussed is seen as useful but not vital (it is clearly useful to preserve mature validation work in written form) and will be prioritized accordingly. Also validations will be turned into acceptance tests where appropriate. Draft workplans at both the LCG and collaboration levels should be completed in a few weeks. A patch to Geant4 V6 was released to address problems found by ATLAS in their pre-production tests. [ATLAS recently reported in the Architects Forum that their simulation validation for Geant4 simulation production is complete, with 500k events simulated and written to POOL.] === FLUKA - Alfredo Ferrari Alfredo not present. FLUKA status and plans were just presented in the last applications area meeting.