We present an implementation for Differential Power Analysis (DPA) that is entirely based on Graphics Processing Units (GPUs). In this paper we make use of advanced techniques offered by the CUDA Framework in order to minimize the runtime. In security testing DPA still plays a major role for the smart card industry and these evaluations require, apart from educationally prepared measurement setups, the analysis of measurements with large amounts of traces and samples, and here time does matter. Most often DPA implementations are tailor-made and adapted to fit certain platforms and hence efficient reference implementations are sparsely seeded. In this work we show that the powerful architecture of graphics cards is well suited to facilitate a DPA implementation, based on the Pearson correlation coefficient, that could serve as a high performant reference, e.g., by analyzing one million traces of 20k samples in less than two minutes.