Simulation Project Leaders Meeting 12-December-2006 ==== Attendees Witek Pokorski (WP), Alberto Ribon (AR), Pere Mato (PM), Rob Veenhof (RV), Gabriele Cosmo (GC). Excused: Alfredo Ferrari (AF), John Harvey (JH). --------------------------------------------------------------------- ==== Minutes - AA Internal Review outcome: new recommendations included in the final report have been acknowledged and will be considered in the planning. --------------------------------------------------------------------- ==== Simulation physics validation - Alberto - Meetings: o At physics validation meeting of October 25th, there was a presentation about performance studies in CMS with Geant4 comparing the different options of multiple-scattering available since release 7.1; another presentation was given by M.Kossov about improvements to the interface between the high-energy quark-gluon string model and CHIPS nuclear fragmentation (QGSC physics list). o At the meeting of November 29th, V.Grichine presented a review of the hadron integral cross-sections in Geant4; V.Uzhinskiy, visitor from Dubna, presented his recent developments in improving the FTF (Fritiof) model in Geant4, which can be considered an alternative physics model for high- energy interactions. The development is still in a 'beta' stage and not yet mature for being tried on calorimeter setups, currently being tested in thin-target setups. This development offering a different approach in the simulation of diffraction, is seen with interest in the context of the studies going on for the hadronic shower shapes, since if applied, it may have an impact in the measurements. There were also two presentations by T.Carli (ATLAS), one about the study on energy response and resolution for the Tile2002 at 90 degrees setup, showing that by adding Bertini cascade model (LHEP_BERT and QGSP_BERT physics lists) improves the measurements. The second presentation was about the unfolding of the test-beam data from digitization, detector noise, ...; the approach, using the original test-beam setup seems to work fine, demonstrating that it is feasible to adopt it in a generic way, by also exporting the detailed geometry through GDML. The approach is being advertised also to other experimental groups in order to allow the reuse of the test-beam data for validation purposes. - Shower shapes: o Some progress has been done so far in the study, it was observed the pi0 production plays an important role; in most Geant4 physics-lists the ratio of total number of pi0 produced is higher than expected (45% instead 33%), and was found that these are coming from the low-energy parameterised component (LEP) used in most physics configurations. This fact seems confirmed by looking at the Bertini model which shows the expected production of pi0 and is the one matching better with the data. This makes QGSP_BERT the best current physics list for reproduction of shower-shape; it describes well data but a the price of about factor of two CPU time penalty. This may indicate that usage of LEP below 10 GeV in the other models should be reviewed. o Currently in the process of writing a note summarizing the studies done in the past months. The note is to be ready by early next year. - 3rd simple benchmark: o Analysis completed with both Fluka and Geant4 official versions. Publication on hold for futher tests required by Fluka experts with a new version of Fluka and new model. o The study will be completed by end of January. - Extension of ATLAS TileCal 2002 analysis with Fluka: o 90 degrees setup analysis currently on hold for corrections in the geometry description and Flugg. o Starting production for the setup with normal incidence angle. Meeting this week with A.Dotti to make the point of the situation. --------------------------------------------------------------------- ==== Generator services - Witek - MCDB: o Planning of the work package to be defined for the next year. Expected a reduction of manpower of 50% in the next 6 months in favor of GENSER. = GENSER: o The draft planning and status report were presented at last GENSER meetings. Most people have shown satisfation with the planning. o Some progress has been already made (O.Zenin, M.Kirsanov), a number of generators packages are already installed in the new AFS area. The implementation of CMT interfaces for the packaging has started and included in the LCG nightly builds for the automatic generation of the packages infrastructure. o Two release cycles are foreseen for GENSER: the installation of the various generators packages (continuous process for new versions / bug fixes...), and a global GENSER release tag, mainly dictated by releases of external packages and evolutions of the testing suite. The first global release of the new GENSER is expected early next year. o The web page has been updated and is still under revision. A Savannah portal for bug reports has been established. The web site will also include soon a page dedicated to testing, where results of the tests on the major MC packages versions will be published and eventually also distributions on relevant physics quantities. - Meetings: o The next monthly Generator Services meeting will be in December 20th. --------------------------------------------------------------------- ==== Fluka - Alfredo - No report. --------------------------------------------------------------------- ==== Simulation framework - Witek - FLUGG: o TileCal studies: not yet at the point to reproduce the geometry of the setup at 90 degrees. Not clear yet the problems are related to the actual geometry description (overlaps of some elements) or in some bug affecting Flugg. More investigation are necessary, eventually by also adopting the new overlap-checking features provided in the most recent releases of Geant4. o There is a clear need of manpower for the maintenance and development of the Flugg tool. - GDML: o The work for the binding GDML->Root has been completed and will be included in the new release of Root. o Improvements in the stand-alone configuration setup of GDML are foreseen, with the implementation of a Configuration script a-la Geant4. B.Lloyd will work on it; he will be asked also to contribute in the extension of GDML for allowing inclusion of user-defined elements to the schema. o Improvements to the documentation are in the plan too. --------------------------------------------------------------------- ==== Geant4 - John (short offline report) - Preparation of developments (including parallel navigation, improved proton elastic scattering, and prototype neutron elastic scattering). - First comparisons with thin target data at 14.6 GeV. The relevant development and testing milestone (scheduled for November 15th) is now completed. - Large scale robustness and regression testing, using local and grid resources, which revealed a number of problems. The key problems identified were addressed in time for the December release. - Geant4 public release 8.2, due December 15th: o Key activity included system integration testing, migration and memory leak testing, preparation of candidate tags and reference tag (made available to experiments two weeks ago). --------------------------------------------------------------------- ==== Garfield - Rob - Garfield activities: o The interface with SRIM is being tested by various groups. o Had contacts recently with an RD50-like group initiating a collaboration for micro-gas detectors. They will require simulation tools, but it is not clear yet which framework will be used. - CMS activation studies with Fluka: o The purpose is to provide a Fluka framework for calculating the induced radioactivity of the various materials in CMS. Such a framework can also be used to compute neutron induced damage to electronics, RPC trigger noise, radiation levels around the experiment etc. So far, the calculations have been done by M.Huhtinen; the results from his analysis are considered accurate. Mika has provided his files and has explained in the detail how he did his calculations. o Two approaches have been considered: 1. from the Geant4 CMS geometry model via Flugg to Fluka 2. native Fluka starting from Mika's geometry description The first approach was thwarted by problems which are probably related to overlaps in the CMS geometry model. It is planned to debug this using the overlap checking features included in the most recent Geant4 releases. Another problem is the questionable chemical composition of some detector elements in the CMS G4 model. The second approach so far was more successful: M.Brugger has converted Mika's files to the Fluka format and has managed to run them through the new Fluka. One of the first steps now will be verifying that the activation calculated with these files matches Mika's numbers. Information about the current layout has been obtained from R.Ranieri and G.Faber. This data is in the process of being translated into Fluka format and will serve to bring the layout up to date. - NA60: o A.Neves is completing his thesis on K0s, Lambda and anti-Lambda production, measuring temperatures and production ratios vs centrality. Atila will start a CERN fellowship in January, at which time the thesis should be essentially ready for defense. --------------------------------------------------------------------- ==== AOB Next meeting scheduled for January 30th.