Argonne National Laboratory

Lessons Learned Implementing User-Level Failure Mitigation in MPICH

TitleLessons Learned Implementing User-Level Failure Mitigation in MPICH
Publication TypeConference Paper
Year of Publication2015
AuthorsBland, W, Lu, H, Seo, S, Balaji, P
Conference NameCCGrid 2015
Date Published05/2015
Other NumbersANL/MCS-P5316-0315
AbstractUser-level failure mitigation (ULFM) is becoming the front-running solution for process fault tolerance in MPI. While not yet adopted into the MPI standard, it is being used by applications and libraries and is being considered by the MPI Forum for future inclusion into MPI itself. In this paper, we introduce an implementation of ULFM in MPICH, a high-performance and widely portable implementation of the MPI standard. We demonstrate that while still a reference implementation, the runtime cost of the new API calls introduced is relatively low.  
PDFhttp://www.mcs.anl.gov/papers/P5316-0315.pdf