Методы разработки параллельных программ на основе машинного обучения
Диссертация
Образец применения методов машинного обучения с использованием характеристик разреженной СЛАУ—работы, авторы которых решают задачу приближенного определения числа обусловленности матрицы СЛАУ на основе алгоритмов машинного обучения. Задача формулируется в виде проблемы классификации, предлагается алгоритм прогнозирования в выбранном пространстве признаков. Результаты обучения и тестирования… Читать ещё >
Список литературы
- Rice J. R., Rosen S. NAPSS. A numerical analysis problem solving system // Proceedings of the 1966 21st national conference. — 1966. — Pp. 51−56.
- Mclnnes L., N orris В., Bhoiumick S., Raghavan P. Adaptive sparse linear solvers for implicit CFD using Newton-Krylov algorithms // 2nd MIT Conference on Computational Fluid and Solid Mechanics, Cambridge, MA (US), 06/17/2003−06/20/2003. 2003.
- Bhowmick S., Raghavan P., Teranishi K. A combinatorial scheme for developing efficient composite solvers // Lecture notes in computer science. — 2004.
- Bhowmick S., Eijkhout V., Freund Y., Fuentes E., Keyes D. Application of machine learning to the selection of sparse linear solvers. — 2006.
- Schonauer W., Hafner H., Weiss R. LINSOL, a parallel iterative linear solver package of generalized CG-type for sparse matrices / / Proc. Eighth SI AM Conference on Parallel Processing for Scientific Computing, Minneapolis, MN, March. 1997.
- Hafner H., Schonauer W., Weiss R. The parallel and portable linear solver package LINSOL // Proceedings of the Fourth European SGI/Cray MPP Workshop, IPP R/46, Max Planck -Institut fur Plasmaphysik, Garching bei Munchen. — 1998.
- George Т., Sarin V. An approach recommender for preconditioned iterative solvers // 2007 International Conference on Preconditioning Techniques for Large Sparse Matrix Problems in Scientific and Industrial Applications. — 2007.
- Pjesivac-Grbovic J., Angskun Т., Bosilca G., Fagg G. E., Gabriel E., Don-garra J. Performance Analysis of MPI Collective Operations // Parallel and
- Distributed Processing Symposium, 2005. Proceedings. 19th IEEE International. — 2005. — Pp. 272a-272a.
- Faraj A., Yuan X., Lowenlhal D. STAR-MPI: self tuned adaptive routines for MPI collective operations // Proceedings of the 20th annual international conference on Supercomputing / ACM Press New York, NY, USA. — 2006. — Pp. 199−208.
- Kuroda H., Katagiri T., Kanada Y. Performance of Automatically Tuned Parallel GMRES (m) Method on Distributed Memory Machines // Hakken Kagaku A05-han. Heisei 11 Nendo. Dai2kai Kaigi Koen Yoshishu. — 1999. — Pp. 11−19.
- Katagiri T., Kuroda II., Kanada Y. A Method for Automatically Tuned Parallel Tridiagonalization on Distributed Memory Vector-Parallel Machines // Vector and Parallel Processing. — 2003.
- Katagiri T., Kise K., Honda II., Yuba T. Effect of auto-tuning with user’s knowledge for numerical software // CF '04: Proceedings of the 1st conference on Computing frontiers. New York, NY, USA: ACM, 2004. — Pp. 12−25.
- Ishii Y., Katagiri T., Honda H. RAO-SS: A Run-time Auto-Tuning Facility for Sparse Solvers with Autopilot / / IP S J SI G Technical Reports. — 2005. — Vol. 2005, no. 19. Pp. 97−102.
- Li X. An overview of SuperLU: Algorithms, implementation, and user interface // ACM Transactions on Mathematical Software (TOMS').— 2005.— Vol. 31, no. 3.-Pp. 302−325.
- Bilmes J., Asanovic K., Chin C. W., DemmelJ. Optimizing Matrix Multiply
- Using PHiPACK: A Portable, High Performance, ANSI C Coding Methodology // Proceedings of International Conference on Supercomputing 1997. -1997. Pp. 340−347.
- Vuduc R., Dernmel J., Bilm. es J. Statistical models for automatic performance tuning // Lecture Notes in Computer Science.— 2001.— Pp. 117 126.
- Vuduc R., Demmel J., Bilmes J. Statistical models for empirical search-based performance tuning // International Journal of High Performance Computing Applications. — 2004. — Vol. 18, no. 1. — P. 65.
- Frigo M., Johnson S. FFTW: an adaptive software architecture for the FFT // Proceedings of the 1998 IEEE International Conference on Acoustics, Speech, and Signal Processing, 1998. ICASSP'98.— 1998. — Vol. 3.
- Whaley R., Dongarra J. Automatically tuned linear algebra software // Proceedings of the 1998 ACM/IEEE conference on Supercomputing (CDROM) / IEEE Computer Society Washington, DC, USA. 1998. — Pp. 1−27.
- Whaley R., Petitet A., Dongarra J. Automated empirical optimizations of software and the ATLAS project // Parallel Computing. — 2001.— Vol. 27, no. 1−2. Pp. 3−35.
- Vuduc R., Demmel J., Yelick K. OSKI: A library of automatically tuned sparse matrix kernels // Journal of Physics: Conference Series. — 2005. — Vol. 16, no. l.-Pp. 521−530.
- Fukui Y., Hasegawa H. Test of Iterative Solvers on ITBL // HPCASIA '05: Proceedings of the Eighth International Conference on High-Performance
- Computing in Asia-Pacific Region. — 2005. — P. 422.
- Cuenca J., Gimenez D., Gonzalez J. Architecture of an automatically tuned linear algebra library // Parallel Computing. — 2004. — Vol. 30, no. 2. — Pp. 187−210.
- Garcia L., Cuenca J., Gimenez. Using Experimental Data to Improve the Performance Modelling of Parallel Linear Algebra Routines // Lecture Notes in Computer Science. 2008. — Vol. 4967. — Pp. 1150−1159.
- Katagiri T., Kise K., Honda H., Yuba T. FIBER: A generalized framework for auto-tuning software // Lecture notes in computer science. — Pp. 146 159.
- Katagiri T., Kanada Y. An efficient implementation of parallel eigenvalue computation for massively parallel processing // Parallel Computing. — 2001. Vol. 27, no. 14. — Pp. 1831−1845.
- Katagiri T., Kise K., Honda H., Yuba T. ABCLibDRSSED: A parallel eigensolver with an auto-tuning facility // Parallel Computing. — 2006. — Vol. 32, no. 3.- Pp. 231−250.
- Xu S., Zhang J. A New Data Mining Approach to Predicting Matrix Condition Numbers // Commun. Inform. Systems. 2004.— Vol. 4, no. 4.— Pp. 325−340.
- Xu S., Lee E. J., Zhang J. An interim analysis report on preconditioners and matrices: Tech. Rep. 388−03: University of Kentucky, Lexington- Department of Computer Science, 2003.
- Self-adapting linear algebra algorithms and software / J. Demmel, J. Don-garra, V. Eijkhout, E. Fuentes, A. Petitet, R. Vuduc, R. Whaley, K. Yelick // Proceedings of the IEEE. 2005. — Vol. 93, no. 2. — Pp. 293−312.
- Eijkhout V., Fuentes E., Eidson T., Dongarra J. The Component Structure of a Self-Adapting Numerical Software System // International Journal of Parallel Programming (IJPP). 2005. — Vol. 33, no. 2−3. — Pp. 137−143.
- Eijkhout V., Fuentes E. A proposed standard for numerical metadata // Innovative Computing Laboratory, University of Tennessee, Tech. Rep. ICL-UT-03−02. 2003.
- Salawdeh I., Cesar E., Morajko A., Margalef Т., buque E. Performance Model for Parallel Mathematical Libraries Based on Historical Knowledgebase // Lecture Notes in Computer Science. — 2008. — Vol. 5168.— Pp. 110−119.
- Automatic performance tuning of parallel mathematical libraries.'
- Davies T. A. Summary of available software for sparse direct methods, http: //cise.uf1.edu/research/sparse/codes.
- Dongarra J., Bullari A. Freely available sofrwarc for linear algebra on the web. — 2009. http://netlib.org/utk/people/JackDongarra/la-sw. html.
- Scientific Application Performance on Candidate PetaScale Platforms / L. Oliker, A. Canning, J. Carter, C. Iancu, M. Lijewski, S. Kamil, J. Shalf et, al. 2007.
- Speck R., Gibbon P., Hoffmann M. Efficiency and scalability of the parallel barnes-hut tree code pepe // Abstract Book of International Conference on Parallel Computations (PARCO-09). 2009.
- Воеводин В. В., Воеводин В. В. Параллельные вычисления.— БХВ-Петербург, 2002. С. 608.
- Wahnig Н., Kotsis G., Haring G. Performance Prediction of Parallel Programs // Messung, Modellierung und Bewertung von. — 1993. — Pp. 64−76.
- Костенко В. А., Трекин А. Г. Генетические алгоритмы решения задач смешанных задач целочисленной и комбинаторной оптимизации при синтезе архитектур ВС // Искуственный интеллект. — 2000. — № 2.
- Davis J., Goadrich М. The Relationship Between Precision-Recall and ROC curves // Proceedings of the 23rd international conference on Machine learning / ACM New York, NY, USA. 2006. — Pp. 233−240.
- Gupta A., Koric S., George T. Sparse Matrix Factorization om Massively Parallel Computers // Proceedings of SC09.— 2009.
- Malony A. D., Mertsiotakis V., Quick A. Automatic scalability analysis of parallel programs based on modeling techniques // Computer Performance
- Evaluation: Modelling Techniques and Tools (LNCS 794) (G. Ilaring and G. Kotsis, eds.). 1994. — Pp. 139−158.
- Mehra P., Schulbach C., Yan J. A comparison of two model-based performance-prediction techniques for message-passing parallel programs // ACM SIGMETRICS Performance Evaluation Review1994, — Vol. 22, no. 1, — Pp. 181−190.
- Вапник В. H. Восстановление зависимостей по эмпирическим данным. — М: Наука, 1979.
- Vapnik V. N. The Nature of Statistical Learning Theory. — Springer, 1995.
- Cortes C., Vapnik V. N. Support-Vector Networks // Machine Learning.— 1995. Vol. 20. Pp. 273−297.
- Smola A., Scholkopf B. A tutorial on support vector regression // Statistics and Computing. 2004. — Vol. 14, no. 3. — Pp. 199−222.
- Schoelkopf, B. and Smola, A. J. Learning with Kernels. — Cambridge, MA: The MIT Press, 2002.
- Schoelkopf B. Statistical Learning and Kernel Methods: Tech. Rep. MSR-TR-2000−23. — Microsoft Corporation, One Microsoft Way, Redmond, WA 98 052: Microsoft Research, 2000.
- Fan R., Chen P., Lin C. Working Set Selection Using Second Order Information for Training Support Vector Machines // The Journal of Machine Learning Research. — 2005. — Vol. 6. — Pp. 1889−1918.
- Chen P., Fan R., Lin C. A Study on SMO-Type Decomposition Methods for Support Vector Machines // IEEE Transactions on Neural Networks. — 2006. Vol. 17, no. 4. — Pp. 893−908.
- Вапник В. H., Червопенкис А. Я. Теория распознавания образов, — М: Наука, 1974. С. 415.
- Akaike H. A new look at the statistical model identification // IEEE transactions on automatic control. — 1974. — Vol. 19, no. 6. — Pp. 716−723.
- Schwarz K. Optimization of the statistical exchange parameter a for the free atoms H through Nb // Physical Review В. — 1972, — Vol. 5, no. 7.— Pp. 2466−2468.
- Stone M. Comments on model selection criteria of Akaike and Schwarz //
- Journal of the Royal Statistical Society. Series В (Methodological). — 1979. — Pp. 276−278.
- Duan K., Keerthi S., Poo A. Evaluation of simple performance measures for tuning SVM hyperparameters // Neurocomputing. — 2003.— Vol. 51, no. 1.- Pp. 41−60.
- Joachims T. Estimating the generalization performance of a SVM efficiently. 1999.
- Evolutionary Feature and Parameter Selection in Support Vector Regression // MICAI 2007, LNAI. 2007. — Vol. 4827. — Pp. 399−408.
- Nachtigal N., Reddy S., Trefethen L. How fast are nonsymmetric matrix iterations? // SIAM Journal on Matrix Analysis and Applications. — 1992. — Vol. 13. P. 778.
- Greenbaum A., Strakos Z. Any nonincreasing convergence curve is possible for GMRES // SIAM Journal of Matrix Analysis Applications. — 1996.
- Chow E., Saad Y. Experimental study of ILU preconditioned for indefinite matrices //J. Comput. Appl. Math. — 1997. — Vol. 86, no. 2. — Pp. 387−414.
- Benzi M., Тита M. A comparative study of sparse approximate inverse pre-conditioners // Appl. Numer. Math. — 1999. — Vol. 30, no. 2−3, — Pp. 305 340.
- Gilbert J. R., Toledo S. An Assessment of Incomplete-LU Preconditioners for Nonsymmetric Linear Systems // Informatica. — 2000. — Vol. 24. — Pp. 409 425.
- Wu K., Milne B. A survey of packages for large linear systems // Lawrence Berkeley National Lab., Rept. LBNL-45U6, Berkeley, CA, Feb. — 2000.
- Kumbhar A., Chakravarthy К., Keshavamurthy R.- Rao G. Utilization of Parallel Solver Libraries to solve Structural and Fluid problems // Tech.
- Rep. Cranes Software International Ltd., Bangalore. — 2005.
- Gould N. I. M., Scott J. A., Hu Y. A numerical evaluation of sparse direct solvers for the solution of large sparse symmetric linear systems of equations // ACM transactions on mathematical software. — 2007.— Vol. 33, no. 2.
- Gupta A., George T., Sarin V. An Experimental Evaluation of Iterative Solvers for Large SPD Systems of Linear Equations: Tech. Rep. RC 24 479. — Yorktown Heights, NY: IBM T. J. Watson Research Center, 2008.
- PETSc Web page / S. Balay, K. Buschelman, W. D. Gropp, D. Kaushik, M. G. Knepley, L. C. Mclnnes, B. F. Smith, H. Zhang // See http://www. mes. anl. gov/petsc. http://www.msc.anl.gov/petsc-2.
- Falgout R. D., Yang U. M. HYP RE: a Library of High Performance Pre-conditioners // Lecture Notes in Computer Science. — 2002, — Vol. 2331.— Pp. 632−641.
- Chang C. C., Lin C. J. — LIBSVM: a library for support vector machines, 2001.— Software available at http://www.csie.ntu.edu.tw/ cjlin/libsvm.
- Whitley D., Rana S., Hechendorn R. B. The island Model Genetic algorithm: On separability, population size and convergence // CIT. Journal of computing and information technology. — 1999. — Vol. 7, no. 1. — Pp. 33−47.
- Combining svms with various feature selection strategies // Technical Report, Department of Computer Science, National Taiwan University. — 2003.
- Chen J., Saad Y. Lanczos Vectors versus Singular Vectors for Effective Dimension Reduction // Knowledge and Data Engineering, IEEE Transaction on. — 2009.- Vol. 21, no. 13.— Pp. 1091−1103. http://www-users.es. umn.edu/~saad/PDF/umsi-2008−02.pdf.
- Davis T. University of Florida sparse matrix collection // NA Digest. — 1997. Vol. 97, no. 23. — P. 7.
- Qian H., Nassif S., Sapatnekar S. Power grid analysis using random walks // Computer-Aided Design of Integrated Circuits and Systems, IEEE Transactions on. 2005. — Vol. 24, no. 8. — Pp. 1204−1224.
- Kozhaya J., Nassif S., Najm F. A multigrid-like technique for power grid analysis // Computer-Aided Design of Integrated Circuits and Systems,
- EE Transactions on. 2002. — Vol. 21, no. 10. — Pp. 1148−1160.
- Chen Т., Chen C. Efficient Large-Scale Power Grid Analysis Based on Preconditioned Krylov-Subspace Iterative Methods // Proc. Design Automation Conference. 2001. — Pp. 559−562.
- Zhao M., Panda R. V., Sapatnekar S. S., Blaauw D. Hierarchical analysis of power distribution networks // Computer-Aided Design of Integrated Circuits and Systems, IEEE Transactions on. — 2002.— Vol. 21, no. 2.— Pp. 159−168.
- Sun K., Zhou Q., Mohanram K., Sorensen D. Parallel domain decomposition for simulation of large-scale power grids // Proceedings of the 2007 IEEE/ACM international conference on Computer-aided design / IEEE Press Piscataway, NJ, USA. 2007. — Pp. 54−59.
- Баландин M. IO., Чепурина Э. П. Методы решения СЛАУ большой размерности.— Новосибирск: Изд-во НГТУ, 2000.
- Schenk О., Gartner К. Solving unsymmetric sparse systems of linear equations with PARDISO // Future Generation Computer Systems. — 2004. — Vol. 20, no. 3. Pp. 475−487.
- Ильин В. П. Методы неполной факторизации для решения алгебраических систем, — М: Наука, Физматлит, 1995.
- Gupta A., Joshi М., Kumar V. WSMP: A high-performance shared and distributed-memory parallel sparse linear equation solver // Report, University of Minnesota and IBM Thomas J. Watson Research Center. — 2001.
- An overview of the Trilinos project / M. A. Heroux, E. T. Phipps, A. G. Salinger, H. K. Thornquist, R. S. Tuminaro, J. M. Willenbring, A. Williams et al. // ACM Transactions on Mathematical Software (TOMS). — 2005.— Vol. 31, no. 3.- Pp. 397−423.
- Попова II. H., Воронов В. Ю., Дэюосан О. В., Медведев М. А. Опыт внедрения современного программного обеспечения на платформе IBM Regatta // Программные системы и инструменты. Тематический сборник N5. — 2004.
- Попова П. Н., Воронов В. Ю., Игумнов В. П., Медведев М. А. Переносимый пакет поддержки распределенной обработки данных с помощью численных методов // Труды Всероссийской научной конференции «Научный Сервис в Сети Интернет-2005». — М: Изд-во МГУ, 2005.
- Специализированная распределенная система обработки экспериментальных данных // Труды второй Всероссийской научной конференции «Методы и средства обработки информации». — М: Изд-во МГУ, 2005. — С. 213−220.
- Воронов В. Ю. Метод автоматического выбора и настройки разреженных решателей СЛАУ // Вестник Московского Университета. Серия 15. Вычислительная математика и кибернетика. — 2009. — Т. 2. — С. 49−56.
- Воронов В. Ю.- Попова H. Н. Моделирование сетей распределения питания СБИС на многоядерном вычислителе // Вычислительные методы и програмлшрова, ние. — 2009. — № 2. — С. 51−60.
- Voronov V. Y., Popova N. N. Parallel Power Grid Simulation on Platforms with Multi Core Processors (acceptcd) // Proceedings of IEEE International Conference on Computing in Engineering, Science and Information (IC-CEIS09). 2009.
- Voronov V. Y., Popova N. N. A Hybrid Simulation of Power Grids using High-Performance Linear Algebra Packages // Abstract book of Numerical Analysis and Scientific Computing with Applications (NASCA-09) conference. 2009. — P. 90.
- Voronov V. Y., Popova N. N. Use of Threaded Numerical Packages for Parallel Power Grid Simulation // Proceedings of International Conference on High Performance Computing, Networking and Communication Systems (HPCNCS-09). 2009. — Pp. 39−45.
- Voronov V. Y., Popova N. N. Automatic Performance Tuning Approach for Parallel Applications Based on Linear Solvers // Abstract Book of International Conference on-Parallel Computations (ParCo-2009). 2009. — P. 29.
- Voronov V. Y. j Popova N. N. Machine Learning Approach to Automatic Performance Tuning of Power Grid Simulator // Abstract Book of 8th European Numerical Mathematics and Advanced Applications (ENUMATH-09) conference. 2009. — P. 291.
- Hernandez V., Roman J., Vidal V. SLEPc: A scalable and flexible toolkit for the solution of eigenvalue problems // ACM Transactions on Mathematical Software (TOMS). 2005. — Vol. 31, no. 3. — Pp. 351−362.
- Lee S. L. Best available bounds for departure from normality // SIMAT.— 1996. Vol. 17. — Pp. 984−991.
- Lee S. Bounds for Departure from Normality and the Frobenius Norm of Matrix Eigenvalues: Tech. rep.: ORNL/TM-12 853, ORNL Oak Ridge National Laboratory (US), 1995.