The GAP project - GPU for realtime applications in high level trigger and medical imaging

The paper describes the GAP Project, whose objective is the deployment of Graphic Processing Units (GPUs) in real-time applications, ranging from trigger selection in high energy physics experiments to medical imaging reconstruction. The final goal of the project is to demonstrate that GPUs have a positive impact in applications that differ for rate, bandwidth, and computational demand, but have a common approach of solving complex problems in real-time using parallel architectures. The relevant aspects under study are the analysis of the system latency, the optimisation of the computational algorithms and the integration with data acquisition systems. As a benchmark application for high-level trigger algorithms in HEP experiments we consider the ATLAS Muon trigger case. In particular we discuss how specific algorithms can be parallelised and thus benefit from the implementation on the GPU architecture, in terms of increased execution speed and more favourable dependency on the complexity of the analysed events. Such improvements are particularly relevant for the foreseen LHC luminosity upgrade where highly selective algorithms will be crucial to maintain a sustainable trigger rate with the many multiple pp interactions per bunch crossing. GPUs can provide a feasible solution also to accelerate the reconstruction of medical images. We discuss the implementation of new computational intense algorithms boosting the performances of Nuclear Magnetic Resonance and Computed Tomography. The deployment of GPUs can significantly reduce the processing time, making it suitable for the use in realtime diagnostic.