Distributed Simulation and Visualization of The ALICE Detector Magnetic Field

TL;DR

This paper presents methods to implement the ALICE detector's magnetic field model in GPU shader language, enabling efficient visualization and integration into particle tracking software.

Contribution

It introduces GPU-based implementations of the magnetic field model, enhancing visualization and computational efficiency for the ALICE detector.

Findings

GPU implementation improves visualization speed.

Model runs efficiently within GPU environment.

Enables new visual analysis of the magnetic field.

Abstract

The ALICE detector at CERN uses properties of the magnetic field acting on charged particles as part of the particle tracking and identification system -- via measuring the strength of bending of charged particles in a magnetic field supplied by electromagnets. During the calibration of the detector the magnetic field was measured by the scientific team. The measurement points were then fitted using Chebyshev polynomials to create a field model. That model is used extensively in particle trajectory reconstruction, but for maximum compatibility it is handled by CPU-only software. We propose multiple approaches to the implementation of the model in a shader language, which allows it to run in a graphical GPU environment. This makes the model run more efficiently, as well as enables it to be used for field visualisation, which was not previously done for ALICE.