Identifying and mitigating the impact of slow-performing nodes in a large supercomputer cluster through the use of machine learning, proxy applications, and scheduling prioritization.
