The International Conference for High Performance Computing, Networking, Storage, and Analysis
The International Conference for High Performance Computing, Networking, Storage, and Analysis
Workshop: XLOOP 2025: The 7th Annual Workshop on Extreme-Scale Experiment-in-the-Loop Computing
Authors: Justin Wozniak (Argonne National Laboratory (ANL)), Sinisa Veseli (Argonne National Laboratory (ANL))
Abstract: Autonomous laboratories streamline exploration of synthesis parameter spaces by automating routine experiments, but their in-line setups depend on high-throughput yet lower-fidelity characterization tools. Integrating these self-driving workflows with user facilities—such as synchrotron and neutron beamlines that provide information-rich, higher-quality data—would minimize wasted trials and accelerate discovery. Connecting autonomous workflows across geographies relies on bi-directional interconnectivity to steer experiments in real time. This paper presents a site-agnostic, framework to stream and analyze data from multiple facilities to enable cross-modal science across scientific user facilities. We demonstrate how beamline experiments at one user facility can be aided by processing experiment data at a different facility, and providing feedback in real time.