An Evaluation of User-Level Failure Mitigation Support in MPI

  • Wesley Bland
  • Aurelien Bouteiller
  • Thomas Herault
  • Joshua Hursey
  • George Bosilca
  • Jack J. Dongarra
Part of the Lecture Notes in Computer Science book series (LNCS, volume 7490)


As the scale of computing platforms becomes increasingly extreme, the requirements for application fault tolerance are increasing as well. Techniques to address this problem by improving the resilience of algorithms have been developed, but they currently receive no support from the programming model, and without such support, they are bound to fail. This paper discusses the failure-free overhead and recovery impact aspects of the User-Level Failure Mitigation proposal presented in the MPI Forum. Experiments demonstrate that fault-aware MPI has little or no impact on performance for a range of applications, and produces satisfactory recovery times when there are failures.


Failure Detection Consensus Algorithm Collective Operation Mean Time Between Failure Alive Process 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Bland, W., Bosilca, G., Bouteiller, A., Herault, T., Dongarra, J.: A proposal for User-Level Failure Mitigation in the MPI-3 standard. Tech. rep., Department of Electrical Engineering and Computer Science, University of Tennessee (2012)Google Scholar
  2. 2.
    Bland, W., Du, P., Bouteiller, A., Herault, T., Bosilca, G., Dongarra, J.: A Checkpoint-on-Failure Protocol for Algorithm-Based Recovery in Standard MPI. In: Kaklamanis, C., Papatheodorou, P., Spirakis, P.G. (eds.) Euro-Par 2012. LNCS, vol. 7484, pp. 477–488. Springer, Heidelberg (2012)Google Scholar
  3. 3.
    Bosilca, G., Bouteiller, A., Brunet, É., Cappello, F., Dongarra, J., Guermouche, A., Hérault, T., Robert, Y., Vivien, F., Zaidouni, D.: Unified Model for Assessing Checkpointing Protocols at Extreme-Scale. Tech. report RR-7950, INRIA (2012)Google Scholar
  4. 4.
    Bougeret, M., Casanova, H., Robert, Y., Vivien, F., Zaidouni, D.: Using group replication for resilience on exascale systems. Tech. Rep. 265, LAWNs (2012)Google Scholar
  5. 5.
    Bouteiller, A., Bosilca, G., Dongarra, J.: Redesigning the message logging model for high performance. CCPE 22(16), 2196–2211 (2010)Google Scholar
  6. 6.
    Buntinas, D., Coti, C., Herault, T., Lemarinier, P., Pilard, L., Rezmerita, A., Rodriguez, E., Cappello, F.: Blocking vs. non-blocking coordinated checkpointing for large-scale fault tolerant MPI protocols. FGCS 24(1), 73–84 (2008)CrossRefGoogle Scholar
  7. 7.
    Cappello, F., Geist, A., Gropp, B., Kalé, L.V., Kramer, B., Snir, M.: Toward exascale resilience. IJHPCA 23(4), 374–388 (2009)Google Scholar
  8. 8.
    Davies, T., Karlsson, C., Liu, H., Ding, C., Chen, Z.: High Performance Linpack Benchmark: A Fault Tolerant Implementation without Checkpointing. In: 25th ICS, pp. 162–171. ACM (2011)Google Scholar
  9. 9.
    Dongarra, J., Beckman, P., et al.: The international exascale software roadmap. IJHPCA 25(11), 3–60 (2011)Google Scholar
  10. 10.
    Du, P., Bouteiller, A., et al.: Algorithm-based fault tolerance for dense matrix factorizations. In: 17th SIGPLAN PPoPP, pp. 225–234. ACM (2012)Google Scholar
  11. 11.
    Fagg, G.E., Dongarra, J.J.: FT-MPI: Fault Tolerant MPI, Supporting Dynamic Applications in a Dynamic World. In: Dongarra, J., Kacsuk, P., Podhorszki, N. (eds.) EuroPVM/MPI 2000. LNCS, vol. 1908, pp. 346–353. Springer, Heidelberg (2000)CrossRefGoogle Scholar
  12. 12.
    Gabriel, E., Fagg, G.E., Bosilca, G., Angskun, T., Dongarra, J., Squyres, J.M., Sahay, V., Kambadur, P., Barrett, B.W., Lumsdaine, A., Castain, R.H., Daniel, D.J., Graham, R.L., Woodall, T.S.: Open MPI: Goals, Concept, and Design of a Next Generation MPI Implementation. In: Kranzlmüller, D., Kacsuk, P., Dongarra, J. (eds.) EuroPVM/MPI 2004. LNCS, vol. 3241, pp. 97–104. Springer, Heidelberg (2004)CrossRefGoogle Scholar
  13. 13.
    Gropp, W., Lusk, E.: Fault tolerance in message passing interface programs. IJHPCA 18, 363–372 (2004)Google Scholar
  14. 14.
    Hadzilacos, V., Toueg, S.: Fault-tolerant broadcasts and related problems. In: Distributed Systems, 2nd edn., pp. 97–145. ACM/Addison-Wesley (1993)Google Scholar
  15. 15.
    Huang, K., Abraham, J.: Algorithm-based fault tolerance for matrix operations. IEEE Transactions on Computers 100(6), 518–528 (1984)CrossRefGoogle Scholar
  16. 16.
    Hursey, J., Graham, R.L., Bronevetsky, G., Buntinas, D., Pritchard, H., Solt, D.G.: Run-Through Stabilization: An MPI Proposal for Process Fault Tolerance. In: Cotronis, Y., Danalis, A., Nikolopoulos, D.S., Dongarra, J. (eds.) EuroMPI 2011. LNCS, vol. 6960, pp. 329–332. Springer, Heidelberg (2011)CrossRefGoogle Scholar
  17. 17.
    Hursey, J., Naughton, T., Vallee, G., Graham, R.L.: A Log-Scaling Fault Tolerant Agreement Algorithm for a Fault Tolerant MPI. In: Cotronis, Y., Danalis, A., Nikolopoulos, D.S., Dongarra, J. (eds.) EuroMPI 2011. LNCS, vol. 6960, pp. 255–263. Springer, Heidelberg (2011)CrossRefGoogle Scholar
  18. 18.
    Lusk, E., Chan, A.: Early Experiments with the OpenMP/MPI Hybrid Programming Model. In: Eigenmann, R., de Supinski, B.R. (eds.) IWOMP 2008. LNCS, vol. 5004, pp. 36–47. Springer, Heidelberg (2008)CrossRefGoogle Scholar
  19. 19.
    Mohan, C., Lindsay, B.: Efficient commit protocols for the tree of processes model of distributed transactions. In: SIGOPS OSR, vol. 19, pp. 40–52. ACM (1985)Google Scholar
  20. 20.
    Sterling, T.: HPC in Phase Change: Towards a New Execution Model. In: Palma, J.M.L.M., Daydé, M., Marques, O., Lopes, J.C. (eds.) VECPAR 2010. LNCS, vol. 6449, p. 31. Springer, Heidelberg (2011)CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  • Wesley Bland
    • 1
  • Aurelien Bouteiller
    • 1
  • Thomas Herault
    • 1
  • Joshua Hursey
    • 2
  • George Bosilca
    • 1
  • Jack J. Dongarra
    • 1
  1. 1.Innovative Computing LaboratoryUniversity of TennesseeUSA
  2. 2.Oak Ridge National LaboratoryUSA

Personalised recommendations