ABSTRACT
Increasing system and algorithmic complexity combined with a growing number of tunable application parameters pose significant challenges for analytical performance modeling. We propose a series of robust techniques to address these challenges. In particular, we apply statistical techniques such as clustering, association, and correlation analysis, to understand the application parameter space better. We construct and compare two classes of effective predictive models: piecewise polynomial regression and artifical neural networks. We compare these techniques with theoretical analyses and experimental results. Overall, both regression and neural networks are accurate with median error rates ranging from 2.2 to 10.5 percent. The comparable accuracy of these models suggest differentiating features will arise from ease of use, transparency, and computational efficiency.
- A. Petit, R. Whaley, J. Dongarra, and A. Cleary. HPL - A portable implementation of the high-performance LINPACK benchmark for distributed-memory computers.Google Scholar
- P.N. Brown, R.D. Falgout, and J.E. Jones. Semicoarsening multigrid on distributed memory machines. SIAM Journal on Scientific Computing, 21(5), 2000. Google ScholarDigital Library
- L. Carrington, A. Snavely, X. Gao, and N. Wolter. A performance prediction framework for scientific applications. In Proceedings International Conference on Computational Science Workshop on Performance Modeling and Analysis (PMA03), June 2003. Google ScholarDigital Library
- L. Carrington, N. Wolter, A. Snavely, and C. Lee. Applying an automatic framework to produce accurate blind performance predictions of full-scale HPC applications. In Department of Defense Users Group Conference, June 2004.Google Scholar
- R. Caruana, S. Lawrence, and C. Giles. Overfitting in neural nets: backpropagation, conjugate gradient, and early stopping. In Neural Information Processing Systems (NIPS), November 2002.Google Scholar
- R. Falgout and U. Yang. HYPRE: A library of high performance preconditioners. In International Conference on Computational Science, 2002. Google ScholarDigital Library
- F. Harrell. Regression modeling strategies. Springer, New York, NY, 2001. Google ScholarDigital Library
- E. Ipek, B. de Supinski, M. Schulz, and S. McKee. An approach to performance prediction for parallel applications. Proceedings of Euro-Par, Springer LNCS, 3648, August 2005. Google ScholarDigital Library
- E. Ipek, S. McKee, B. de Supinski M. Schulz, and R. Caruana. Efficiently exploring architectural design spaces via predictive modeling. In Proceedings Architectural Support for Programming Languages and Operating Systems (ASPLOS XII), October 2006. Google ScholarDigital Library
- P. Joseph, K. Vaswani, and M.J. Thazhuthaveetil. Construction and use of linear regression models for processor performance analysis. In Proceedings of the 12th Symposium on High Performance Computer Architecture, Austin, Texas, February 2006.Google ScholarCross Ref
- D. Kerbyson, H. Alme, A. Hoisie, F. Petrini, H. Wasserman, and M. Gittings. Predictive performance and scalability modeling of a large-scale application. In Proceedings IEEE/ACM Supercomputing, November 2001. Google ScholarDigital Library
- B. Lee and D. Brooks. Accurate and efficient regression modeling for microarchitectural regression models. In Proceedings International Symposium on High-Performance Computer Architecture (HPCA-13), February 2007.Google Scholar
- G. Marin and J. Mellor-Crummey. Cross-architecture performance predictions for scientific applications using parameterized models. In Proceedings International Conference on Measurement and Modeling of Computer Systems (Sigmetrics), June 2004. Google ScholarDigital Library
- T. Mitchell. Machine Learning. WCB/McGraw Hill, 1997. Google ScholarDigital Library
- M. Riedmiller and H. Braun. A direct adaptive method for faster backpropagation learning: The RPROP algorithm. In Proceedings IEEE International Conference on Neural Networks, May 1993.Google ScholarCross Ref
- C. Stone. Comment: Generalized additive models. Statistical Science, 1:312--314, 1986.Google ScholarCross Ref
- T. Yang, X. Ma, and F. Mueller. Cross-platform performance prediction of parallel applications using partial execution. In Proceedings IEEE/ACM Supercomputing, November 2005. Google ScholarDigital Library
- J. Yi, D. Lilja, and D. Hawkins. Improving computer architecture simulation methodology by adding statistical rigor. IEEE Computer, November 2005. Google ScholarDigital Library
- A. Zell and et. al. SNNS: Stuttgart neural network simulator, user manual, version 4.2. In User Manual, Version 4.2, University of Stuttgart.Google Scholar
Index Terms
- Methods of inference and learning for performance modeling of parallel applications
Recommendations
Connectionist models for approximate solutions of non-linear equations in one variable
In this paper, six neural network models for the computation of an approximate real root of a given non-linear equation are proposed. The models are recurrent with one or more layers. The delay and the feedbacks are automatically taken care by the ...
Extreme learning machine and its applications
Recently, a novel learning algorithm for single-hidden-layer feedforward neural networks (SLFNs) named extreme learning machine (ELM) was proposed by Huang et al. The essence of ELM is that the learning parameters of hidden nodes, including input ...
A comparative predictive analysis of neural networks (NNs), nonlinear regression and classification and regression tree (CART) models
Numerous articles comparing performances of statistical and Neural Networks (NNs) models are available in the literature, however, very few involved Classification and Regression Tree (CART) models in their comparative studies. We perform a three-way ...
Comments