Simulation Project Leaders Meeting 21-February-2012 ==== Attendees John Apostolakis (JA), Andrea Dotti (AD), Pere Mato (PM), Witold Pokorski (WP), Rob Veenhof (RV), Gabriele Cosmo (GC). Excused: Alfredo Ferrari (AF) --------------------------------------------------------------------- ==== Minutes --------------------------------------------------------------------- ==== Simulation physics validation & framework - Andrea - Meetings o The activity of the Physics Validation project will be presented at the MC4BSM workshop in March at Cornell University. The goal is to exchange expertise with the generators experts on how the validation of MC is performed for LHC experiments. Hopefully synergies will be found. - Activities o New validation plots have been added to the project web-page. Plots are extracted from the LPCC workshop presentation. o Validation of the Geant4 release 9.5 has been performed on the GRID. A report CERN-LCGAPP-2011-04 has been produced. The same test suite is now used also for monthly releases. o LHCb has developed a stand-alone application to test Multiple Scattering models. This is a simplified setup, similar in the spirit to the SimplifiedCalorimeter application. It is our intention to use this code, that we received from our LHCb colleagues, to test this aspect of simulations on a periodic basis. --------------------------------------------------------------------- ==== Generator services - Witek - GENSER: o 10 new versions of generators introduced: epos-1.99.crmc.r2790.p1, libtests-1.02, pythia8-160, photos++-3.3, tauola++-1.0.5.p1, rivet-1.7.0, pythia6-426.2, vincia-1.0.27, herwig++-2.5.2, thepeg-1.7.2. o Anton Pytel has finished on January; his last task was on GDML, where he has rewritten the Root-GDML interface in Root from Python to C++ and provided several fixes, also through cross import/export files with Geant4. - HepMC: o Three new releases announced and consequent three new global releases of the generators stack; main feature is the addition of CMake scripts in HepMC and fixing some bugs on units. --------------------------------------------------------------------- ==== Fluka - Alfredo - No report. --------------------------------------------------------------------- ==== Geant4 - John, Gabriele - Releases: o In addition to release 9.5 announced on December 2nd and the first prototype of a multi-threaded version of Geant4 (Geant4-MT, based on release 9.4.p01) announced in November, a new patch release 9.4.p03 was made on December 13th. Among all corrections included in the patch to note a fix resolving the long-standing problem of tracks stuck at boundaries in presence of a magnetic field, reported by ATLAS; a fix for small steps at high energy due to bad scattering in the Urban multiple-scattering; a fix for smoothing of cross-sections of Bremsstrahlung, responsible for a discrepancy observed by CMS in lateral shower shapes. o Release 9.4.p03 is now candidate for production of 2012 Monte Carlo in ATLAS and CMS. CMS is also considering to port on top of 9.4.p03 the new Bremsstrahlung process introduced in release 9.5, which provides improved cross-sections derived from the original sources; this helps improving on the parameterisations used in Geant3 and in the previous Geant4 Bremsstrahlung process, which had imprecision of up to 5% for some elements and energy values. - Developments: o First tests are being made with an interface to the UrQMD event generator (version 1.3 cr). This provides new capabilities for ion-ion interactions up to large energies ( TeV ), and can be used also for any hadron-hadron interactions except hyperons. o Improved triton-nucleus cross-sections are being prepared. o The testing for verifying event reproducibility has restarted. An increased incidence of problems is observed compared to the old release 9.3 when restarting with an intermediate event. Under investigation. o Moved to using CDash for continuous integration of proposed developments and fixes. The system integration testing is now shared with other Geant4 collaborators from different Institutions. - Experiments/users feedback: o An increase in memory use in Geant4 9.5 was identified when simulating low energy neutrons using the neutron-HP module. Among the reasons identified to date, the increase of the number of isotopes now available in the revised G4NDL library, and a code change in HP elastic which, once fixed helps reducing the overhead by ~40%. Fixes will be included in a future patch to 9.5. o A rare problem of big energy non-conservation in 9.4.p01 was reported by ATLAS and affecting anti-proton annihilation in CHIPS. This has been found to occur every 7-8 million interactions for C, N, O. A work-around has been prepared and is being tested. - Meetings: o A first Geant4 steering-board meeting has been held on February 3rd, where the process for the election of the new working-group coordinators was defined. Another meeting will take place on the 24th focussed on the preparation of the overall work-plan for 2012. o The next Technical Forum will take place at CERN on March 27th to discuss the 2012 work-plan and define the terms of support for release 9.4. --------------------------------------------------------------------- ==== Garfield - Rob - No report --------------------------------------------------------------------- ==== AOB Next meeting scheduled for April 24th.