Simulation Project Leaders Meeting 25-September-2012 ==== Attendees (virtual meeting) John Apostolakis (JA), Andrea Dotti (AD), Alfredo Ferrari (AF), Witold Pokorski (WP), Rob Veenhof (RV), Gabriele Cosmo (GC). --------------------------------------------------------------------- ==== Minutes --------------------------------------------------------------------- ==== Simulation physics validation & framework - Andrea, Witek - Validation: o The FNAL validation team will be involved in further develop the existing infrastructure to dynamically chose a configuration (a few versions of Geant4 + selected physics lists), and generate corresponding histograms. The timescale however is still not clear. Also to be evaluated if to use the CDash/CTest infrastructure to systematically run the validation application for producing the required data files. o The GRID validation suite (Simplified-Calorimeter) has been greatly improved by A.Dotti, in terms of automation and integration with CERNVM. The lack of long term support of the underlying Diane tool still remains an open problem. --------------------------------------------------------------------- ==== Generator services - Witek - GENSER: o New generators installed: winhac-1.31, baurmc-1.0, pyquen-1.5.1, powheg-box-r2092, epos-1.99.crmc.v3200, mctester-1.25.0a, photos++-3.51, amcatnlo-1.0, pythia6-426p12, mctester-1.25.0, sherpa-1.4.1.2, epos-1.99.crmc.v3130, lhapdf-5.8.8, rivet-1.8.1.fj3, mcfm-6.2b evtgen-1.0.0. o Started discussing with B.Hegner from SFT/SPI about a direct cooperation of GENSER and the SPI/external software. Planning to introduce changes in GENSER both on the technical side (to adopt the build system used by SPI) as well as on the organizational side (to directly involve Benedikt in the coordination of GENSER). --------------------------------------------------------------------- ==== Fluka - Alfredo - No report. --------------------------------------------------------------------- ==== Geant4 - John, Gabriele - Releases: o The 9.6-beta release was made on schedule on June 29th, 2012. First feedback has been received from users. To be noted that new data sets G4EMLOW-6.27 and G4NDL-4.1 are required. A new version of the CLHEP library, CLHEP-2.1.2.3, is recommended and is used internally (depending on the installation configuration). A degradation of about 5% in CPU time was found in release 9.6-beta. First improvements to recover were made and recovered 1/3 of this overhead; further work is underway. o A new release of the multi-threading prototype, Geant4-MT, based on the latest release 9.5.p01 was released on August 15th. Its configuration and build procedure uses Cmake exclusively. It incorporates improvements for creating applications and fixes to ensure the use of the customized thread-private allocators provided in Geant4-MT. - Development: o Part of the development for release 9.6 and already included in the Beta release is a series of improvements in the EM physics builders used in HEP applications. This includes the adoption of the Wentzel multiple scattering model for electrons (only at energies above 100 MeV) and for all other particles (spanning all energies). This enables the improved treatment of large angle scattering at high energies. Configuration to use the relativistic e+/e- pair production model at energies above 80 GeV (to correct deviation of cross-section found in existing model in this limit). Fixes and improvements in the angular generators for Bremsstrahlung and photo-electric. Configuration currently under test has been changed to use generator developed by V.Grichine. The performance improvement in the Livermore Rayleigh scattering model now allow its use also in HEP applications, if needed. It is used already in the most accurate EM builders, including "Opt3". o Physics-lists in preparation for Geant4 9.6: decisions were taken for the configuration of production physics lists in 9.6. They covered the key lists: FTFP_BERT (recommended), QGSP_FTFP_BERT and QGSP_BERT (legacy). All have been or will be migrated to use the new Bertini/FTFP stopping models for stopping particles, lepto-nuclear and electro-nuclear interactions, in place of CHIPS models. The cross-sections of electro-nuclear, gamma-nuclear and quasi-elastic which have been extracted from CHIPS will be used. For electromagnetics, the configuration of EM builders will be revised: Opt1 (for homogeneous calorimeters) is refined to coincide with the options chosen together with CMS; Opt2 (tuned for vertex reconstruction) is targeted to address the needs and parallel fully the choices of LHCb (it is configured with multiple scattering with higher precision of displacement for high energy particles and less reduced precision for showers in calorimeters); a new choice, Opt4, further refines the precision choice (Opt3) in order to give the best combined standard/lowenergy models. Other choices, which have not been changed significantly, are: the 'default' Opt0 (used by ATLAS) no Rayleigh, with accurate treatment of multiple-scattering for calorimeter showering; Opt3 is the most accurate electron and gamma transport that still uses primarily the CPU-performant 'standard' models (configured with stricter step limits, for applications sensitive to transport of lower energy particles); it is used in studies for HEP detectors, and in applications for radiotherapy, imaging and space (based on options from the 'standard' plus Rayleigh from Livermore). - Meetings: o The annual Geant4 Collaboration Meeting has been held in Chartres (France) on September 10-14th. Plans have been established for Geant4 Release 10 in 2013, in which multithreading will be incorporated in the base Geant4 source code. It will be possible to choose at installation time between the multi-thread capable version and the sequential version. The beta, development release for this is scheduled for June 2013. There have been reports on the exploration of GPU for Geant4 at FNAL, CERN and the joint KEK/SLAC/Stanford-ICME efforts. FNAL and CERN will collaborate closely in exploring HEP-targeted GPU acceleration. Progress was reported in the comparison of Geant4 electromagnetic module with EGSnrc and Penelope. Good overall agreement was reported, with some discrepancies which are currently attributed to different treatment of multiple-scattering. o The next Geant4 workshop will be held in Sevilla (Spain) in September 2013. o The 9th Geant4 Space Users' Workshop will take place in Barcelona (Spain) on March 4-6th. --------------------------------------------------------------------- ==== Garfield - Rob - Main activities: o Focussing on GEMs development, with emphasis on topics relevant for the ALICE and CMS upgrades. o Regular software maintenance. - Meetings & Lectures: o RD51 mini-week in June. o Visit of Steve Biagi in mid-July. o RD51 collaboration meeting in Bursa. New meeting in Stony Brook in early October. o INFN school in Firenze, were 2 lectures were given. o School in Mainz (1 lecture) o Schools in Bodrum (3 lectures) - Students projects: o Student from Mexico as NMA during the summer, worked on photon detection using thick GEMs, report due in the RD51 October meeting. - Publications: o Papers on photon detection (MHSP) and neBEM have been accepted. o Paper on GEM still in progress. --------------------------------------------------------------------- ==== AOB Next meeting scheduled for December 11th.