Simulation Project Leaders Meeting 13-April-2010 ==== Attendees John Apostolakis (JA), Alfredo Ferrari (AF), Alberto Ribon (AR), Witold Pokorski (WP), Rob Veenhof (RV), Gabriele Cosmo (GC). Excused: Pere Mato. --------------------------------------------------------------------- ==== Minutes --------------------------------------------------------------------- ==== Simulation physics validation & framework - Alberto - Validation studies: o At the 27th Geant4 Technical Forum taken on March 30th, there has been a presentation by A.Zhemchugov for the HARP-CDP group related to the comparison of proton, pi+, pi- beam on beryllium, copper and tantalum with Geant4-9.3, using QGSP_BERT, QGSP_BIC and FTFP_BERT physics-lists, with momenta between 1.5, 15 Gev/C between 20-50 and 50-125 degrees. QGSP_BIC shows the worst results with clear discontinuities at 10 GeV; QGSP_BERT marks below the data and also shows discontinuity at 10 GeV and in the number of secondaries; FTFP_BERT shows smoother behavior with values mostly above data by 20-50%, eventually requiring some tuning. It will be asked to collect the data in numerical form to use for tuning. o Physics Validation meeting on March 31st: S.Banerjee presented the validation studies carried on the CMS-HCAL using Geant4-9.3, but still using an older version of CHIPS model and no Fritiof-based physics-lists; it is observed a CPU performance improvement of roughly 10%. Presentation also by M.Simonyan on ATLAS-Tilecal at 90 degrees, where the physics-list QGSP-BIC shows the best pi- and p response, but bad longitudinal and later shower shapes and marked discontinuity in the energy response as a function of the beam energy. Also tested CHIPS which looks interesting for longitudinal and lateral profiles, and quite smooth as expected; still showing too high response and too narrow energy resolution. Best performing are the Fritiof-based lists, in particular FTFP_BERT but also QGSP_FTFP_BERT (where Fritiof replaces the parameterised model in QGSP_BERT). Also presentation by D.Ward from CALICE, testing Geant4 9.3 hadronics in the ECAL (Si-W), mainly looking at the beginning of the hadronic shower. The analysis considers the following observables: fraction of events that do not interact in the ECAL, visible energy for those interacting the ECAL, transverse and longitudinal shower profile with respect to the first interaction. Most lists describe reasonably well the data within 10-20%, best performing are FTFP_BERT and FTF_BIC. - Meetings: o Next meeting is scheduled around mid-May. --------------------------------------------------------------------- ==== Generator services - Witek - Meetings: o MC4LHC workshop held on March 29th, focussed on readiness for LHC experiments; also given a presentation on the status of GENSER. - GENSER: o New versions of Sherpa, LHAPDF, Herwig++. New version of Rivet. - Validation: o Implemented few validation tests using Rivet with Pythia8 and Herwig++: still using data from CDF but with the purpose to soon try out the new LHC data. o Also adopting the HepMC analysis tool, for histograms comparing the different generators. Work carried out in collaboration with the Desy group. - HepMC: o Announced Beta release of HepMC 2.06 and now waiting for feedback. --------------------------------------------------------------------- ==== Fluka - Alfredo - Meetings: o Planning advanced course and workshop in Portugal the first week of October close to Lisbon: will be structured with half days tutorials and half day presentations by users. - Fluka: o New features being developed and almost ready for geometry viewer and debugger (overlaps detection); also allowing for simple adaptation of the geometry setup graphically from screen. o Planning a new release Fluka-2010.1 (or .2) likely after the summer; it will contain quite a lot of new features and will fully support gfortran 64-bits native. - Announced drop of support for the VMC interfaces, its use will be no longer permitted or supported outside ALICE. A detailed review revealed intrinsic weaknesses in the interface, unsupported functionalities and features, serious bugs, hard-coded defaults hard to customize or modify in a flexible way, no possibility to control tracking in magnetic field, etc... All deficiencies preventing the user from exploiting the full capabilities of the code and leading in many cases to inaccurate or even wrong results in the simulation. The Fluka coordination committee therefore decided to drop support and officially informed all users about this decision. - Started activity for the Envision European project for simulation tools for monitoring hadron-therapy; got some funding for this activity which is carried together with the Milano group. - Finalized and signed commercial agreement with Siemens for use in hadron-therapy treatment systems. Some funding expected from this for work on hadron-therapy. - Automatic coupling of the code with accelerator tracking codes SixTrack, to improve jobs setup with Fluka. Still in optimization stage; coupling tool controlling processes running in independent nodes. - CNGS neutrino facility: currently taking data; hope to have ICARUS first data to test soon. --------------------------------------------------------------------- ==== Geant4 - John - Releases: o Preparing new patch 9.3.p01 for release 9.3, including retuning of CHIPS fragmentation; fixes for ion processes; fixes in high-energy parameterised model for k0 inelastic; fixes for Windows/VC++. o New release 9.4-beta planned for end of June. Final release 9.4 scheduled for December. - Developments: o M.Kossov presented at the last Physics Validation meeting the improvements for neutron elastic scattering in CHIPS, with emphasis on energies down to MeV, isotopes-wise; improved elastic cross-sections for kaons, anti-protons, lambdas also developed. o Fixes for numerical issues identified in FTFP and Bertini. Fix in pre-compound model for light targets (Al). o Improved the infrastructure for running regression suite, for runs locally and on the GRID. Added new observables (moments). - Architecture review: o Ongoing meetings, during March. Considering key interfaces for improvement related to performance and multi-threading. Some areas identified for improvement or investigation include physics tables, use of memory in geometry/tracking. - Meetings: o TF meeting: feedback included requests to improve kaon cross-sections by LHCb; improved visualisation of subtracted solid (CMS, repeat); development of tracking for magnetic monopoles (CMS). Technical requests from LHCb on more uniform way for labeling warnings and errors to aid in extracting; consistent order of arguments in constructors for physics lists' builders. CMS report CPU improvement of about 5% from 9.2 to 9.3; will move production to 9.3 soon. o Geant4 Collaboration Workshop at ESA/ESTEC in October 4-9th. --------------------------------------------------------------------- ==== Garfield - Rob - Meetings: o VCI 2010 in Vienna: numerous talks on LHC startup; reduced pressure on detector developers as a result of the more relaxed LHC calendar; surge in work related to medical imaging. o RD51 miniweek at CERN. o VECC Medical imagine (Kolkata): various high-resolution, high-efficiency detectors were shown, the Roeschtigraben remains. - Garfield: o neBEM: several Green's functions added, weighting fields taken care of, first iteration of space charge and surface charge (more to come). o Magboltz: updates of the excitation cross sections of Ar, Xe, Ne, He and H now complete. o Heed: interfacing work starting. - Measurements: o Orsay: excitation transfer rate measurements in Ar/CO2 (for inclusion in a microscopic gas gain model), discussions planned for next week with the Turkish group, measurements to follow. o Saclay: gain fluctuation measurements (for validation of models) to be continued in 2 weeks. --------------------------------------------------------------------- ==== AOB Next meeting scheduled for June 1st.