A GPU-Based Kalman Filter for Track Fitting

Abstract Computing centres, including those used to process High-Energy Physics data and simulations, are increasingly providing significant fractions of their computing resources through hardware architectures other than x86 CPUs, with GPUs being a common alternative. can provide excellent computational performance at good price point for tasks that be suitably parallelized. Charged particle (track) reconstruction is computationally expensive component HEP reconstruction, thus needs use available in an efficient way. In this paper, implementation Kalman filter-based track fitting using CUDA running on presented. This utilizes the ACTS (A Common Tracking Software) toolkit; open source experiment-independent toolkit reconstruction. The details parallelization approach described, along specific challenges such implementation. Detailed benchmarking results discussed, which show encouraging gains over CPU-based representative configurations. Finally, perspective future directions these studies outlined. These include more complex realistic scenarios studied, anticipated developments software frameworks standards may up possibilities greater flexibility improved performance.