When fitting models to data with low statistics and near physical boundaries, additional measures need to be taken to ensure proper statistical coverage. This is especially true in neutrino physics, where neutrino-matter weak interactions are rare. NOvA, a long-baseline neutrino oscillation experiment, aims to measure the parameters governing neutrino oscillations within the PMNS oscillation model. In order to ensure proper statistical coverage, NOvA uses the computationally intensive Feldman-Cousins (FC) procedure. The FC procedure involves fitting thousands of independent pseudoexperiments to generate empirical statistical distributions that are then used to calculate the significance of measurements. Here, work done by NOvA and SciDAC collaborators to implement the FC procedure utilizing the High Performance Computing facilities at NERSC, is presented. With these tools, NOvA’s complete FC corrections for their recent neutrino + antineutrino appearance analysis were processed in 36 hours: a speedup factor of 50 when compared to the methods used in previous analyses.
