Implementation-Oblivious Transparent Checkpoint-Restart for MPI
arXiv:2309.14996
Abstract
This work presents experience with traditional use cases of checkpointing on a novel platform. A single codebase (MANA) transparently checkpoints production workloads for major available MPI implementations: "develop once, run everywhere". The new platform enables application developers to compile their application against any of the available standards-compliant MPI implementations, and test each MPI implementation according to performance or other features.
17 pages, 4 figures