Simulation Project Leaders Meeting 9-September-2008 ==== Attendees John Apostolakis (JA), Alfredo Ferrari (AF), Alberto Ribon (AR), Pere Mato (PM), Rob Veenhof (RV), Gabriele Cosmo (GC). --------------------------------------------------------------------- ==== Minutes - A request for a technical student has been advanced for work on the Geant4 geometry and maintenance/development on Flugg. --------------------------------------------------------------------- ==== Simulation physics validation & framework - Alberto - Meetings: o The next meeting is scheduled for September 24th, where a presentation on the ATLAS TileCal 90-degree analysis is planned; the analysis is essentially complete and a paper is under preparation. - Test beam analyses: o Work is still in progress for the ATLAS combined calorimeter. Final results are expected soon. - Experiments feedback: o Some concern has been expressed in ATLAS about the presence of discontinuities in the simulation resulting in a lack of smoothness in the transition region between different overlapping models. Particularly, for the QGSP_BERT physics list, a discontinuity is observed between 9 and 10 GeV, in the transition between LEP and Bertini; for FTF_BIC and FTFP_BERT between 4 and 5 GeV. Currently trying to reproduce this phenomenon by using simplified setups, with the objective of improving matching of models in the physics-lists. --------------------------------------------------------------------- ==== Generator services - Alberto - Meetings: o The last meeting was held on June 25th and was dedicated to Sherpa; the package is growing and getting more and more complete with the addition of several new features. It is now used more heavily by experiments, which also provide quite positive feedback. o The next Generator Services meeting will be on October 1st and will be dedicated to MC-Tester, the tool for comparison of decays in the generators. - GENSER: o Several new versions of generator packages have been installed, see the Generators Services web page for reference: http://lcgapp.cern.ch/project/simu/generator/ o Progress has been made in the general bootstrap script, following also the good feedback received from the Durham group. The system build infrastructure is getting extended as new versions of the generators are added. O.Zenin is the main person working on it. o MC@NLO: a discussion with the author and ATLAS people took place to clarify the need for reviewing the package infrastructure; it turned out that for the time being, the current distribution strategy for the package is satisfactory. - HepMC: o The transition of LHCb to HepMC-2 has been completed, thanks also to the work by L.Sonnenschein and P.Canal; now able to read data generated with old HepMC versions; ATLAS and LHCb now use the same version of HepMC. - MCDB: o The migration to a new server has been completed. --------------------------------------------------------------------- ==== Fluka - Alfredo - Releases: o A new release of Fluka (2008.04) is planned within this month. New features include the new neutron library, which has been now tested extensively; it improves neutron thermalisation. The new release also will include the first release of a neutrino- nucleus generator. o The nuclear data-base (NNDC radiation decay DB) in the code has been updated. The need for an update of the DB has been triggered by the observation of wrong results for a couple of isotopes, following the development for the introduction of electron conversion lines. - Meetings: o A training course is planned in Paris at the Nuclear Agency of Energy at the end of September. They will also start distributing the Fluka neutron library. o The course held at CERN at the end of June saw the participation of almost 30 people; to be noted the successful organisation and preparation of PCs. It is not clear yet if there will be one or two courses in 2009, it will mostly depend on the availability of people involved. - Development: o Still busy on calculations for the machine related to damage to the electronics. Quite a lot of soft electronics already planned for being relocated at the next shutdown, some already relocated. The CNGS exercise was very useful in this sense. o A CERN contract for safety studies (activation calculations) in ITER has been assigned in form of a fellowship. o Going to switch the web-site with new look in the next weeks. Comments on the prototype page (http://www.fluka.org/fluka.php) are welcome. --------------------------------------------------------------------- ==== Geant4 - John - Releases: o A patch is being prepared for 9.1 (9.1.p03). Most of the fixes are also included in 9.2-beta, plus some more, including a fix in the field propagation causing a rare crash in ATLAS (about 2 per million events). 9.1.p03 also contains corrections for two issues reported by the HARP-CDP group last March (see Physics Validation meeting of June 18th): the missing Fermi motion in the high energy quasi-elastic modeling (first introduced in Geant4 8.3) and the incorrect sampling of p_t in the parameterised (LEP, Gheisha derived) model used in several physics lists (LHEP, QGSP) for 8-9 GeV protons and pions. - Feedback from experiments: o ATLAS has reported great stability of their simulation based on 8.3.p02 (one failure every 500K events); the cause for that rare failure has been identified in the field propagation and the fix has recently been provided (it will also be included in 9.1.p03). - Meetings: o The preparation for the Geant4 Collaboration Workshop in October in Kobe (Japan) is going on; 59 people registered so far, with users coming both from Europe and Japan. It is scheduled users talks on LHC, space and medicine applications; 2.5 days will be dedicated to users workshop and 2.5 days as Collaboration restricted meeting. The program is almost finalised. An Oversight Board meeting will take place in Kobe as well. This is the 2nd Collaboration Workshop held in Japan, the first one being in Niigata 10 years ago. o The last Technical Forum was held on July 2nd, at CERN. It included summaries of the experiments' feedback related to their current production runs, and open issues from test beam validation relating to physics modeling. o A meeting was held last week at LAPP, bringing together developers in the electro-magnetic working groups, focusing on design convergence and future work items. o The last EUDET annual meeting will take place at NIKHEF, Amsterdam in October 6-8th. V.Uzhinskiy will participate for the VALSIM work package on Geant4. o Presentations about Geant4 developments will take place at the next IEEE NSS Conference in Dresden. - Manpower: o The spokesperson elections in Geant4, were held in July. John Apostolakis has been reelected spokesperson of the Collaboration, for a term of 2 years. --------------------------------------------------------------------- ==== Garfield - Rob - Interfaces: o Currently writing C++ classes to make Magboltz, Garfield and Heed callable from Geant4 and Root. Stefan Guindon, non-memberstate summer student from Canada, has used these to make a demonstration model in which Geant4 tracks a muon up to an Atlas MDT tube, then leaves Heed generate ionisation electrons which Garfield tracks through the tube using Magboltz electron transport tables. Although the simulation is essentially complete (signal calculation and electronics simulation were not done yet, but these are not believed to be potential stumbling blocks), several areas of work remain. In particular, the interface needs to be extended to a broader range of geometries. Also, a number of coding issues need to be ironed out. - Development: o Tracking of electrons at the molecular level has been extended to include avalanches. Work is ongoing to compare the results of this technique with the Legler and Alkhazov models and to understand whether this technique reproduces experimental data. Experimental determination of Penning transfer probabilities by now includes about 15 gas mixtures at a range of pressures. - Meetings: o RD51 meeting foreseen October 13th-15th in Paris. --------------------------------------------------------------------- ==== AOB Next meeting scheduled for November 4th.