Neutrinos are electrically neutral, weakly interacting fundamental particles with the surprising property that they change between different types, or “flavors”, as they travel. This flavor-changing phenomenon has been measured at different distances (baselines) over the past few decades, but several anomalous measurements at “short” baselines of about a kilometer remain in tension with theoretical predictions. If confirmed, these anomalies would imply a new fundamental particle and would radically change our understanding of particle physics. 

The Short-Baseline Neutrino (SBN) program, now operating at Fermilab, studies neutrinos using multiple advanced liquid argon time projection chamber (LArTPC) detectors placed in a neutrino beam and is poised to collect the world’s largest and most detailed data set of neutrino interactions to date. This data set will definitively resolve the short-baseline anomaly and simultaneously enable a rich program of neutrino-nucleus interaction research. The goal of this ALCC award is to integrate the high-performance computing (HPC) resources at the Argonne Leadership Computing Facility into the SBN research program to accelerate and enhance its scientific output. The HPC resources provided by the award will enable GPU-accelerated machine learning algorithms to be used at scale in the analysis of the many petabytes of data generated by the SBN experiments. This unique capability will drive the development of new algorithms by facilitating fast, iterative reprocessing of SBN data, and will enable the production of the large-scale simulated data sets necessary for making precision measurements. 

This award will strengthen collaboration between national laboratories while showcasing the advantages of using HPC resources for a multi-stage particle physics research program. In addition to the scientific output of the SBN program, this award will result in a model for real-time data transfer between national labs, coordinated and scalable workflow execution, and multiple processing pipelines for real and simulated data sets. Further, the timely physics results and computational methods of the SBN program, enabled by this ALCC award, will inform the physics program and computational infrastructure of the next-generation flagship neutrino experiment: the Deep Underground Neutrino Experiment (DUNE).