Volume-based Representation of the Magnetic Field

Simulation and reconstruction of events in high-energy experiments require the knowledge of the value of the magnetic field at any point within the detector. The way this information is extracted from the actual map of the magnetic field and served to simulation and reconstruction applications has a large impact on accuracy and performance in terms of speed. As an example, the CMS high level trigger performs on-line tracking of muons within the magnet yoke, where the field is discontinuous and largely inhomogeneous. In this case the high level trigger execution time is dominated by the time needed to access the magnetic field map.For this reason, an optimized approach for the access to the CMS field was developed, based on a dedicated representation of thedetector geometry. The detector is modeled in terms of volumes, constructed in such a way that their boundaries correspond to the fiel d discontinuities due to changes in the magnetic permeability of the materials. The field within each volume is therefore continuous. When the field for a given point in space is requested, the corresponding volume is found using an optimized algorithm that exploits explicitly the layout and the symmetries of the CMS detector. The value of the field within the volume is then obtained by interpolation from a regular grid of the values resulting from a simulation with TOSCA or, when it is available, from a parameterization. The main clients of the magnetic field, which are the simulation (GEANT) and the propagation of track parameters and errors in the reconstruction, can be made aware of the magnetic field volumes by connecting the per-volume magnetic field providers to the corresponding volume in the respective geometries. In this way the global volume search is by-passed and the access to the field is sped up significantly.