Analyzing time-dimension communication characterizations for representative scientific applications on supercomputer systems

Exascale computing is one of the major challenges of this decade, and several studies have shown that communications are becoming one of the bottlenecks for scaling parallel applications. The analysis on the characteristics of communications can effectively aid to improve the performance of scientific applications. In this paper, we focus on the statistical regularity in time-dimension communication characteristics for representative scientific applications on supercomputer systems, and then prove that the distribution of communication-event intervals has a power-law decay, which is common in scientific interests and human activities. We verify the distribution of communication-event intervals has really a power-law decay on the Tianhe-2 supercomputer, and also on the other six parallel systems with three different network topologies and two routing policies. In order to do a quantitative study on the power-law distribution, we exploit two groups of statistics: bursty vs. memory and periodicity vs. dispersion. Our results indicate that the communication events show a “strong-bursty and weak-memory” characteristic and the communication event intervals show the periodicity and the dispersion. Finally, our research provides an insight into the relationship between communication optimizations and time-dimension communication characteristics.

[1]  R. Sah,et al.  Market Shares: Some Power Law Results and Observations , 2003 .

[2]  George Kingsley Zipf,et al.  Human behavior and the principle of least effort , 1949 .

[3]  Russell C. Coile,et al.  Lotka's frequency distribution of scientific productivity , 1977, J. Am. Soc. Inf. Sci..

[4]  Canqun Yang,et al.  MilkyWay-2 supercomputer: system and application , 2014, Frontiers of Computer Science.

[5]  E. Lu,et al.  Avalanches and the Distribution of Solar Flares , 1991 .

[6]  Aleksandar Jemcov,et al.  OpenFOAM: A C++ Library for Complex Physics Simulations , 2007 .

[7]  Lada A. Adamic,et al.  The Nature of Markets in the World Wide Web , 1999 .

[8]  Azer Bestavros,et al.  Self-similarity in World Wide Web traffic: evidence and possible causes , 1996, SIGMETRICS '96.

[9]  D J PRICE,et al.  NETWORKS OF SCIENTIFIC PAPERS. , 1965, Science.

[10]  S.M. Kay,et al.  Spectrum analysis—A modern perspective , 1981, Proceedings of the IEEE.

[11]  Chen Cui,et al.  Detailed and clock-driven simulation for HPC interconnection network , 2016, Frontiers of Computer Science.

[12]  Mark E. J. Newman,et al.  Power-Law Distributions in Empirical Data , 2007, SIAM Rev..

[13]  Constantinos E. Goutis,et al.  A methodology for speeding up loop kernels by exploiting the software information and the memory architecture , 2015, Comput. Lang. Syst. Struct..

[14]  C. Svaneborg.  Large-scale Atomic/Molecular Massively Parallel Simulator , 2011 .

[15]  Yuhua Tang,et al.  Reducing Static Energy in Supercomputer Interconnection Networks Using Topology-Aware Partitioning , 2016, IEEE Transactions on Computers.

[16]  Quentin F. Stout,et al.  The Use of the MPI Communication Library in the NAS Parallel Benchmarks , 1999 .

[17]  D. Zanette,et al.  Vertical transmission of culture and the distribution of family names. , 2000, nlin/0009046.

[18]  G. Neukum,et al.  Crater Size Distributions and Impact Probabilities on Earth from Lunar, Terrestrial Planeta, and Asteroid Cratering Data , 1994 .

[19]  Raymond A. K. Cox,et al.  The concentration of commercial success in popular music: An analysis of the distribution of gold records , 1994 .

[20]  D. Altman,et al.  Statistics notes: Measurement error , 1996 .

[21]  G. Yule,et al.  Some Statistics of Evolution and Geographical Distribution in Plants and Animals, and their Significance. , 1922, Nature.

[22]  Peter V. Coveney,et al.  Large-scale molecular dynamics simulation of DNA: implementation and validation of the AMBER98 force field in LAMMPS , 2004, Philosophical Transactions of the Royal Society of London. Series A: Mathematical, Physical and Engineering Sciences.

[23]  Anthony Skjellum,et al.  Using MPI - portable parallel programming with the message-parsing interface , 1994 .

[24]  Christina Freytag,et al.  Using Mpi Portable Parallel Programming With The Message Passing Interface , 2016 .

[25]  Adilson E. Motter,et al.  A Poissonian explanation for heavy tails in e-mail communication , 2008, Proceedings of the National Academy of Sciences.

[26]  Michalis Faloutsos,et al.  On power-law relationships of the Internet topology , 1999, SIGCOMM '99.

[27]  Ibm Blue,et al.  Overview of the IBM Blue Gene/P Project , 2008, IBM J. Res. Dev..

[28]  Michael Mitzenmacher,et al.  A Brief History of Generative Models for Power Law and Lognormal Distributions , 2004, Internet Math..

[29]  Wenguang Chen,et al.  PHANTOM: predicting performance of parallel applications on large-scale parallel machines using a single node , 2010, PPoPP '10.

[30]  Martin Schulz,et al.  Visualizing Network Traffic to Understand the Performance of Massively Parallel Simulations , 2012, IEEE Transactions on Visualization and Computer Graphics.

[31]  Jack J. Dongarra,et al.  The LINPACK Benchmark: past, present and future , 2003, Concurr. Comput. Pract. Exp..

[32]  R. Hidalgo,et al.  Conditions for the emergence of scaling in the inter-event time of uncorrelated and seasonal systems , 2006 .

[33]  D. Turcotte,et al.  Fractality and Self-Organized Criticality of Wars , 1998 .

[34]  Azer Bestavros,et al.  Self-similarity in World Wide Web traffic: evidence and possible causes , 1997, TNET.

[35]  Charles Gide,et al.  Cours d'économie politique , 1911 .

[36]  J. A. Morgan,et al.  Calculation of the Residual Sum of Squares for all Possible Regressions , 1972 .

[37]  B. Gutenberg,et al.  Frequency of Earthquakes in California , 1944, Nature.

[38]  David H. Bailey,et al.  The Nas Parallel Benchmarks , 1991, Int. J. High Perform. Comput. Appl..

[39]  D. Legates,et al.  Evaluating the use of “goodness‐of‐fit” Measures in hydrologic and hydroclimatic model validation , 1999 .

[40]  Fabio Checconi,et al.  Characterization of the Communication Patterns of Scientific Applications on Blue Gene/P , 2011, 2011 IEEE International Symposium on Parallel and Distributed Processing Workshops and Phd Forum.

[41]  Yuetsu Kodama,et al.  Efficient MPI Collective Operations for Clusters in Long-and-Fast Networks , 2006, 2006 IEEE International Conference on Cluster Computing.

[42]  Keqin Li,et al.  Communication Analysis and Performance Prediction of Parallel Applications on Large-Scale Machines , 2016 .

[43]  Rajeev Kohli,et al.  Some Empirical Regularities in Market Shares , 2006, Manag. Sci..

[44]  Robert A. van Engelen,et al.  Efficient Symbolic Analysis for Optimizing Compilers , 2001, CC.

[45]  D. van der Spoel,et al.  GROMACS: A message-passing parallel molecular dynamics implementation , 1995 .

[46]  H. Simon,et al.  ON A CLASS OF SKEW DISTRIBUTION FUNCTIONS , 1955 .

[47]  Xuejun Yang,et al.  An effective speedup metric for measuring productivity in large-scale parallel computer systems , 2009, The Journal of Supercomputing.

[48]  V. Latora,et al.  Complex networks: Structure and dynamics , 2006 .

[49]  J. Rodgers,et al.  Thirteen ways to look at the correlation coefficient , 1988 .

[50]  Zhi-Dan Zhao,et al.  Relative clock verifies endogenous bursts of human dynamics , 2012 .

[51]  M. E. J. Newman,et al.  Power laws, Pareto distributions and Zipf's law , 2005 .

[52]  Juan Chen,et al.  Time-Dimension Communication Characterization of Representative Scientific Applications on Tianhe-2 , 2015, 2015 IEEE 17th International Conference on High Performance Computing and Communications, 2015 IEEE 7th International Symposium on Cyberspace Safety and Security, and 2015 IEEE 12th International Conference on Embedded Software and Systems.

[53]  V. Pareto,et al.  Vilfredo Pareto. Cours d’Économie Politique , 1897 .

[54]  Kwang-Il Goh,et al.  Burstiness and memory in complex systems , 2006 .

[55]  Al Geist,et al.  Major Computer Science Challenges At Exascale , 2009, Int. J. High Perform. Comput. Appl..

[56]  廖湘科,et al.  High Performance Interconnect Network for Tianhe System , 2015 .

[57]  H. Raghav Rao,et al.  A parallel hypercube algorithm for discrete resource allocation problems , 2006, IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans.

[58]  Xin Yuan,et al.  LFTI: A New Performance Metric for Assessing Interconnect Designs for Extreme-Scale HPC Systems , 2014, 2014 IEEE 28th International Parallel and Distributed Processing Symposium.

[59]  S. Franz,et al.  Critical Phenomena in Natural Sciences: Chaos, Fractals, Selforganization and Disorder: Concepts and Tools , 2004 .

[60]  George Woodbury,et al.  An Introduction to Statistics , 2001 .

[61]  William J. Dally,et al.  Principles and Practices of Interconnection Networks , 2004 .

[62]  M. Newman.  Power laws, Pareto distributions and Zipf's law , 2005 .