PCcheck: Persistent Concurrent Checkpointing for ML
Year: 2025
Type: conference-paper
Abstract: Training large-scale machine learning (ML) models is expensive and time-intensive, consuming many hardware accelerators for days or weeks. As the scale of hardware deployments and training time continue to grow, the probability of failures also increases. The desire to use cheaper cloud resources, such as spot VMs, to lower costs also dramatically increases the frequency of failures. The standard approach to deal with failures is to periodically pause training and checkpoint model parameters to persistent storage. Unfortunately, today's checkpointing mechanisms introduce high overhead when applied at high frequencies, yet frequent checkpointing is necessary to avoid long recovery times.
Authors: Foteini Strati, Michal Friedman, Ana Klimovic
Institution: ETH Zurich
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4407216788