Evaluating User-Level Fault Tolerance for MPI Applications
Ignacio Laguna, David F. Richards, Todd Gamblin, Martin Schulz, Bronis R. de Supinski · 2014
The User Level Failure Mitigation (ULFM) interface has been proposed to provide fault-tolerant semantics in MPI. Previous work has presented performance evaluations of the interface; yet questions related to its programability and applicability remain unanswered. In this paper, we present our experiences on using ULFM in a case study (a large molecular dynamics application) to shed light on the advantages and difficulties of this interface to program fault-tolerant MPI applications. We found that, although ULFM is suitable for applications with work-decomposition flexibility (e.g., master-slave), it provides few benefits for more general (e.g., bulk synchronous) MPI applications.