|Title||Lessons Learned Implementing User-Level Failure Mitigation in MPICH |
|Publication Type||Conference Paper |
|Year of Publication||2015 |
|Authors||Bland, W, Lu, H, Seo, S, Balaji, P |
|Conference Name||CCGrid 2015 |
|Date Published||05/2015 |
|Other Numbers||ANL/MCS-P5316-0315 |
|Abstract||User-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.