CDF detector simulation framework and performance

TL;DR

The paper presents a GEANT3-based detector simulation framework integrated into the AC++ environment for the CDF Run II experiment. It emphasizes a modular, generic-programming design that unifies geometry, digitizers, and MC data while enabling run-time configurability and seamless integration with reconstruction. The authors validate the framework across silicon, COT, ToF, muon systems, CLC, and calorimeters, achieving good agreement with collider data and demonstrating substantial speedups via GFLASH for calorimeter simulation. An upgrade path to GEANT4 via a low-level API is discussed, underscoring the framework's flexibility and scalability for large-scale Monte Carlo production.

Abstract

The CDF detector simulation framework is integrated into an AC++ application used to process events in the CDF experiment. The simulation framework is based on the GEANT3 package. It holds the detector element geometry descriptions, allows configuration of digitizers at run-time and manages the generated data. The design is based on generic programming which allows for easy extension of the simulation framework. The overall design, details of specific detector components and in particular the performance of the CDF simulation compared to collider data are described.

Figures (14)

• Figure 1: Generator related modules within the AC++ framework.
• Figure 2: Generator related modules within the AC++ framework.
• Figure 3: Cluster length (left) and cluster profile (right) for the axial side (top) and stereo side (bottom) of Layer 1 of SVX II. Data are the black points, Monte Carlo are the shaded (red) points.
• Figure 4: Unbiased axial residuals (Layer 2). Date are the points with error bars; MC is the histogram.
• Figure 5: Track residuals (hit displacement minus track displacement) for data (points) and MC (histogram).