Reasons for a Pessimistic or Optimistic Message Logging Protocol in MPI Uncoordinated Failure Recovery - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2009

Reasons for a Pessimistic or Optimistic Message Logging Protocol in MPI Uncoordinated Failure Recovery

Abstract

With the growing scale of high performance computing platforms, fault tolerance has become a major issue. Among the various approaches for providing fault tolerance to MPI applications, message logging has been proved to tolerate higher failure rate. However, this advantage comes at the expense of a higher overhead on communications, due to latency intrusive logging of events to a stable storage. Previous work proposed and evaluated several protocols relaxing the synchronicity of event logging to moderate this overhead. Recently, the model of message logging has been refined to better match the reality of high performance network cards, where message receptions are decomposed in multiple interdependent events. According to this new model, deterministic and non-deterministic events are clearly discriminated, reducing the overhead induced by message logging. In this paper we compare, experimentally, a pessimistic and an optimistic message logging protocol, using this new model and implemented in the Open MPI library. Although pessimistic and optimistic message logging are, respectively, the most and less synchronous message logging paradigms, experiments show that most of the time their performance is comparable.
No file

Dates and versions

inria-00424017 , version 1 (13-10-2009)

Identifiers

  • HAL Id : inria-00424017 , version 1

Cite

Aurélien Bouteiller, Thomas Ropars, George Bosilca, Christine Morin, Jack Dongarra. Reasons for a Pessimistic or Optimistic Message Logging Protocol in MPI Uncoordinated Failure Recovery. IEEE International Conference on Cluster Computing (Cluster 2009), Sep 2009, New Orleans, United States. ⟨inria-00424017⟩
172 View
0 Download

Share

Gmail Facebook X LinkedIn More