Simulation Project Leaders Meeting 7-April-2009 ==== Attendees John Apostolakis (JA), Alberto Ribon (AR), Gabriele Cosmo (GC), Pere Mato (PM). Excused: Rob Veenhof (RV). --------------------------------------------------------------------- ==== Minutes --------------------------------------------------------------------- ==== Simulation physics validation & framework - Alberto - Meetings: o At the physics validation meeting held on April 1st, a presentation by M.Simonyan from ATLAS focussed on comparison between results obtained with releases 9.1 and 9.2 of Geant4 for the ATLAS TileCal test-beam analysis at 90 degrees. QGS physics lists are reported to be quite stable, while expected changes are observed in Fritiof, with shorter and narrower showers in 9.2 in overall worst agreement with the data. Should be noted that significant changes to the FTF model have been included in 9.2, based on thin-target data; in particular the tuning done against the HARP data will have to be reviewed. The main concern is in any case related to the known discontinuities observed in the visible energy, due to the transition between models, problem which remains unresolved; energy transition with Fritiof-based physics lists is anyhow smoother compared to other physics lists. The variation in the energy scale (electrons) has been confirmed to be no bigger of 1%. o The next meeting is scheduled for May 13th, where it is also planned a presentation by the ILC/CALICE group. - The main activity currently going on is concerning the transition between models and some progress is being made. M.Kossov is also working to extend the CHIPS model in the pre-compound region below the pion threshold. --------------------------------------------------------------------- ==== Generator services - Alberto - Meetings: o The next generator services meeting will be on April 8th and will be dedicated to MCDB and its status in CMS, currently used as database to store LHE files (i.e. ascii parton-level files in the Les Houches Event format). - GENSER: o The new release of HepMC 2.05 is a bit late in the schedule but is expected soon. o Various new versions of supported generators have been requested by the experiments and already included. o Received also a request to make available MacOSX installations for the main generators. It will be discussed at the next planning meeting. o Currently working on the porting to SLC5; starting with one flavor, 32-bit gcc-4.3.2 and already completed for LHPDF, Pythia6, Herwig, Jimmy, Photos, Hijing, HydJet and Pythia8; working on Herwig++, while some problems have been experienced with Tauola and reported to the authors. Once complete this first phase and dropped SLC4, SLC5 64-bits builds will be added as well. o Autotools: a version of Pythia6 was provided but no feedback from the experiments has been received yet. It is proposed to progress with the transition anyhow also for the remaining packages; the strategy will be discussed at the next planning meeting scheduled at the end of May. --------------------------------------------------------------------- ==== Fluka - Alfredo - No much activity in the code is going on right now. Busy with calculation for the shielding and relocation of the electronics in the tunnel from critical areas. - A new minor release 2008.3b.0 of Fluka has been made available last week; Compton of individual electronic level is now released plus also some bug fixes. - Still working on finalising the implementation of damage predictions of displacement per atom (DPA); the work about the energy prediction for non-ionising recoils is almost complete, still something to do for tiny recoils. Once ready, time will be also required for testing before getting the new development in the production version. --------------------------------------------------------------------- ==== Geant4 - John - Releases: o A new patch to release 9.2 has been released on March 16th, Geant4 9.2.p01, including most of the fixes reported since December. Fixes corrected the computation of the emission probability in de_excitation model, the addition of smearing of Coulomb barriers for d, t, he3 and alpha particles, and the probability for the emission of light ions in the pre-compound model; tuned the absorption coefficient in Bertini cascade; activated deletion of processes, models and cross-sections at the end of the job, to improve the memory allocation. o A first prototype multi-threaded Geant4, based on release 9.1.p01, has been created and is being benchmarked on different systems (8-24cores). A prototype multi-threaded version based on 9.2 is planned for June 2009. - Developments: o Working on extending the GRID testing to make use of US sites (OSG); facing some challenges due to the differences of tools and disk space available on sites. - Feedback from experiments: o ATLAS moving to 9.2 provisionally, pending validation; reporting an increase of memory usage for the initialization of about 25%, potentially due to physics tables. o LHCb reports 15% jobs failure for events stuck all on the same volume; no additional and detailed information has been provided yet. - Meetings: o A meeting was held last week with the CALICE collaboration at DESY. CALICE is using Geant4 for their simulation, and have run a number of test-beams at CERN and Fermilab. The information provided by their studies is very useful, since complementary to LHC test beams. Their highly granular calorimeter allows to study the first interaction at some level detail. In addition the evolution of a hadronic shower can be studied both event by event and in space with additional information, such as the lateral profile along a shower's longitudinal evolution. They expressed their interest to use FLUKA, if they can access it within their existing simulation framework. The potential for using a customized version of the VMC was discussed. o The final report for the internal Delta review has been provided and is positive in noting significant progress and in providing new recommendations. The planning of future developments presented was favorably received, and the need for additional resources for some areas including documentation was noted. o A Geant4 course is being organised and expected to take place at CERN in September 21-25th. --------------------------------------------------------------------- ==== Garfield - Rob - No report --------------------------------------------------------------------- ==== AOB Next meeting scheduled for May 26th.