Nondeterminism is an increasingly entrenched property of high-performance computing (HPC) applications and has recently been shown to seriously hamper debugging and reproducibility efforts. Tools for addressing the nondeterministic debugging problem have emerged, but they do not provide methods for systematically cataloging the nondeterminism in a given application. We propose a three-phase workflow for representing executions of nondeterministic message passing interface programs as event graphs, quantifying their structural similarity with graph kernels, and applying machine learning techniques to investigate shared properties across applications. We present an empirical study comparing two graph kernels’ suitability for this task and propose future uses of the methodology.
AhnDHLeeGLGopalakrishnanG, et al. (2013) Overcoming extreme-scale reproducibility challenges through a unified, targeted, and multilevel toolset. In: Proceedings of the First International Workshop on Software Engineering for high Performance Computing in Computational Science and Engineering, Denver, Colorado, USA, 17–21 November 2013, pp. 41–44. ACM.
2.
ChappD (2017) Study of the Impact of Nondeterminism on Numerical Reproducibility and Debugging at the Exascale. MS Thesis, University of Delaware, Newark, DE.
3.
ChappDJohnstonTTauferM (2015) On the need for reproducible numerical accuracy through intelligent runtime selection of reduction algorithms at the extreme scale. In: 2015 IEEE International Conference on Cluster Computing (CLUSTER) (ed Bilof R), Chicago, IL, USA, 8–11 September 2015, pp. 166–175. IEEE.
4.
ChappDSatoKAhnDH, et al. (2018) Record-and-replay techniques for HPC systems: a survey. Supercomputing Frontiers and Innovations5(1): 11–30.
5.
ClevelandMABrunnerTAGentileNA, et al. (2013) Obtaining identical results with double precision global accuracy on different numbers of processors in parallel particle Monte Carlo simulations. Journal of Computational Physics251: 223–236.
6.
GhoshSDasNGoniçalvesT, et al. (2018) The journey of graph kernels through two decades. Computer Science Review27: 88–111.
KranzlmüllerDGrabnerSVolkertJ (1996) Event graph visualization for debugging large applications. In: Proceedings of the SIGMETRICS Symposium on Parallel and Distributed Tools, Philadelphia, PA, USA, 22–23 May 1996, pp. 108–117. New York, NY, USA: ACM.
9.
LamportL (1978) Time, clocks, and the ordering of events in a distributed system. Communications of the ACM21(7): 558–565.
10.
PreisslRSchulzMKranzlmüllerD, et al. (2008) Using MPI communication patterns to guide source code transformations. In: BubakMvan AlbadaGDDongarraJSlootPMA (eds) International Conference on Computational Science—ICCS 2008. ICCS 2008. Lecture Notes in Computer Science, Vol. 5103, Kraków, Poland, 23–25 June 2008, pp. 253–260. Berlin, Heidelberg: Springer
11.
RamonJGärtnerT (2003) Expressivity versus efficiency of graph kernels. In: Proceedings of the First International Workshop on Mining Graphs, Trees and Sequences, pp. 65–74.
12.
RichardsDBrantleyPDawsonS, et al. (2016) Quicksilver (No. Quicksilver; 004652WKSTN00). Livermore, CA (United States): Lawrence Livermore National Laboratory (LLNL).
13.
RodriguesAFVoskuilenGRHammondSD, et al. (2016) Structural Simulation Toolkit (SST) (No. SAND2016-3693PE). Albuquerque, NM (United States): Sandia National Lab.(SNL-NM).
14.
SatoKAhnDHLagunaI, et al. (2015) Clock delta compression for scalable order-replay of non-deterministic parallel applications. In: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis, Austin, TX, 15–20 November 2015, p. 62. New York, NY, USA: ACM.
15.
SatoKAhnDHLagunaI, et al. (2017) Noise injection techniques to expose subtle and unintended message races. In: ACM SIGPLAN Notices, Vol. 52, No. 8, Austin, TX, USA, 04–08 February 2017, pp. 89–101. New York, NY, USA: ACM.
16.
SeoEKhanMMHMohapatraP, et al. (2011) Exposing complex bug-triggering conditions in distributed systems via graph mining. In: 2011 International Conference on Parallel Processing (ICPP), Taipei, Taiwan, 13–16 September 2011, pp. 186–195. Washington, DC, USA: IEEE.
17.
ShervashidzeNSchweitzerPLeeuwenEJV, et al. (2011) Weisfeiler-lehman graph kernels. Journal of Machine Learning Research12: 2539–2561.
18.
TauferMAndersonDPCicottiP, et al. (2005) Homogeneous redundancy: a technique to ensure integrity of molecular simulation results using public computing. In: Proceedings of the 19th International Parallel and Distributed Processing Symposium, Denver, CO, USA, 4–8 April 2005, pp. 1–8. IEEE.
19.
VishwanathanSVNSchraudolphNNKondorR, et al. (2010) Graph kernels. Journal of Machine Learning Research11: 1201–1242.
