Status of the BaBar Simulation Geant4 Workshop 30 September 2002 - PowerPoint PPT Presentation

Status of the BaBar Simulation Geant4 Workshop 30 September 2002 Dennis Wright

Outline BaBar Overview Comparison projects – EM process validation – Hadronic process validation Performance

The BaBar Detector

Overview of the BaBar Simulation Event Generator Detector Response Reconstruction And Background And Analysis Mixing (Bear) (SimApp) Tracking , Physics , Hit Scoring ( Bogus , GEANT4 ) Objectivity Database

Use of Geant4 in BaBar BaBar uses: BaBar does not use: – Geometry – Standard tracking/navigation – Hit-scoring – Detector response – Decay processes – Persistence – EM physics processes (< 10 GeV) – Low energy hadronic processes (< 10 GeV)

BaBar/Geant4 Validation (1) 1 million events generated by October 2000 – B0B0bar, Bhabha, dimuons Use to test gross features of simulation: – Material model validation – Tracking/resolution tests – G3/G4/Data comparison of EM processes – Rough check of hadronic processes – Tests of performance/robustness

BaBar/Geant4 Validation (2) Detailed comparisons of MC and data – 11 million event sample completed July 2002 • > 25 different event types simulated – Goals: • fine tuning of EM processes (check recent improvements) • Take a serious look at hadronic processes • Examine calorimetry, tracking, PID

EM processes/calorimetry

EM processes/calorimetry

EM processes/calorimetry

Tracking

Tracking

Hadronic Validation Currently using low energy parameterized model – Not especially appropriate for BaBar energies (50 MeV – 10 GeV) – Several fixes reported by LCD group at SLAC – passed to G4 – Some included in toolkit, some to be included soon Use latest MC data set and BaBar “thin target” tests to search for and fix other problems Very recent results from tests –hadronic group hasn’t seen them yet

BaBar “Thin Target” Hadronic Tests

Hadronic Validation Hadron kinetic model (40 MeV – 10 GeV) will be perfect for BaBar – Will switch to that as soon as it is ready SLAC (T. Koi, D. Wright) has recently joined in the hadron kinetic model validation effort First effort – develop validation test bench for p,n, pi, K – Compare data, kinetic model at 500 MeV – few GeV

Performance (1) Simulation stage of generic B0-B0bar event includes event generator, tracking, hit-scoring – On 866 MHz PIII takes 5.0 s/evt – Used Geant4 4.0 Currently running MC production at 19 sites (1 billion events so far) Run failures due to Geant4 getting rare – < 1 per million events – A few annoyances remain

Performance (2) BaBar recently decided to double the MC/data ratio (to 3) Speed issues now more important than ever Geant4 part of BaBar simulation time 30-40% Options: – Simplify detector geometry – Use parameterized showers – Use faster machines (new 1.4 GHz machines being tested)

Conclusions EM validation well in hand – A little tweaking remains Hadronic validation beginning in earnest – Results will be passed to hadronics group Huge amount of MC data to be produced using Geant4 over life of experiment – Seeking ways to speed up simulation