A Large-Grained Parallel Algorithm for Nonlinear Eigenvalue Problems Using Complex Contour Integration Takeshi Amako, Yusaku Yamamoto and Shao-Liang Zhang.

Slides:



Advertisements
Similar presentations
Chapter 0 Review of Algebra.
Advertisements

Chapter 6 Matrix Algebra.
Answering Approximate Queries over Autonomous Web Databases Xiangfu Meng, Z. M. Ma, and Li Yan College of Information Science and Engineering, Northeastern.
DEPARTMENT OF CIVIL ENGINEERING THE NATIONAL UNIVERSITY OF SINGAPORE Euro-SiBRAM2002 International Colloquium Prague - Czech Republic June 24 to 26, 2002.
Copyright 2011, Data Mining Research Laboratory Fast Sparse Matrix-Vector Multiplication on GPUs: Implications for Graph Mining Xintian Yang, Srinivasan.
Copyright © Cengage Learning. All rights reserved.
Ch 7.6: Complex Eigenvalues
Chapter 6 Equations 6.1 Solving Trigonometric Equations 6.2 More on Trigonometric Equations 6.3 Trigonometric Equations Involving Multiples Angles 6.4.
1 Department of Civil and Environmental Engineering Sungkyunkwan University 비대칭 박벽보의 개선된 해석이론 및 방법 An Improved Theory and Analysis Procedures of Nonsymmetric.
ECE 530 – Analysis Techniques for Large-Scale Electrical Systems Prof. Hao Zhu Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign.
Lecture 5 Newton-Raphson Method
Applied Linear Algebra - in honor of Hans SchneiderMay 25, 2010 A Look-Back Technique of Restart for the GMRES(m) Method Akira IMAKURA † Tomohiro SOGABE.
A Linearization method for Polynomial Eigenvalue Problems using a contour integral Junko Asakura, Tetsuya Sakurai, Hiroto Tadano Department of Computer.
Lecture 17 Introduction to Eigenvalue Problems
Linear Transformations
Motion Analysis Slides are from RPI Registration Class.
Chapter 7 Reading on Moment Calculation. Time Moments of Impulse Response h(t) Definition of moments i-th moment Note that m 1 = Elmore delay when h(t)
Efficient Methodologies for Reliability Based Design Optimization
Development of Empirical Models From Process Data
An introduction to iterative projection methods Eigenvalue problems Luiza Bondar the 23 rd of November th Seminar.
Parallel Implementation of the Inversion of Polynomial Matrices Alina Solovyova-Vincent March 26, 2003 A thesis submitted in partial fulfillment of the.
Monica Garika Chandana Guduru. METHODS TO SOLVE LINEAR SYSTEMS Direct methods Gaussian elimination method LU method for factorization Simplex method of.
CISE-301: Numerical Methods Topic 1: Introduction to Numerical Methods and Taylor Series Lectures 1-4: KFUPM.
MATH 685/ CSI 700/ OR 682 Lecture Notes Lecture 6. Eigenvalue problems.
Component Reliability Analysis
Exercise problems for students taking the Programming Parallel Computers course. Janusz Kowalik Piotr Arlukowicz Tadeusz Puzniakowski Informatics Institute.
An approach for solving the Helmholtz Equation on heterogeneous platforms An approach for solving the Helmholtz Equation on heterogeneous platforms G.
1 Hybrid methods for solving large-scale parameter estimation problems Carlos A. Quintero 1 Miguel Argáez 1 Hector Klie 2 Leticia Velázquez 1 Mary Wheeler.
Algorithms for a large sparse nonlinear eigenvalue problem Yusaku Yamamoto Dept. of Computational Science & Engineering Nagoya University.
CISE-301: Numerical Methods Topic 1: Introduction to Numerical Methods and Taylor Series Lectures 1-4: KFUPM CISE301_Topic1.
Least-Mean-Square Training of Cluster-Weighted-Modeling National Taiwan University Department of Computer Science and Information Engineering.
AEGIS 2008 Annual Assembly 11 December 2008 PALESS – Parallel Analogue and Logical Simulation System Marko Dimitrijević Faculty of electronic engeneering,
Efficient Integration of Large Stiff Systems of ODEs Using Exponential Integrators M. Tokman, M. Tokman, University of California, Merced 2 hrs 1.5 hrs.
PDCS 2007 November 20, 2007 Accelerating the Complex Hessenberg QR Algorithm with the CSX600 Floating-Point Coprocessor Yusaku Yamamoto 1 Takafumi Miyata.
Self Organization of a Massive Document Collection Advisor : Dr. Hsu Graduate : Sheng-Hsuan Wang Author : Teuvo Kohonen et al.
In-Won Lee, Professor, PE In-Won Lee, Professor, PE Structural Dynamics & Vibration Control Lab. Structural Dynamics & Vibration Control Lab. Korea Advanced.
© 2011 Autodesk Freely licensed for use by educational institutions. Reuse and changes require a note indicating that content has been modified from the.
Accelerating the Singular Value Decomposition of Rectangular Matrices with the CSX600 and the Integrable SVD September 7, 2007 PaCT-2007, Pereslavl-Zalessky.
On the Use of Sparse Direct Solver in a Projection Method for Generalized Eigenvalue Problems Using Numerical Integration Takamitsu Watanabe and Yusaku.
1 Complex Images k’k’ k”k” k0k0 -k0-k0 branch cut   k 0 pole C1C1 C0C0 from the Sommerfeld identity, the complex exponentials must be a function.
* 김 만철, 정 형조, 박 선규, 이 인원 * 김 만철, 정 형조, 박 선규, 이 인원 구조동역학 및 진동제어 연구실 구조동역학 및 진동제어 연구실 한국과학기술원 토목공학과 중복 또는 근접 고유치를 갖는 비비례 감쇠 구조물의 자유진동 해석 1998 한국전산구조공학회 가을.
Parallel Solution of the Poisson Problem Using MPI
A Flexible New Technique for Camera Calibration Zhengyou Zhang Sung Huh CSPS 643 Individual Presentation 1 February 25,
EECS 274 Computer Vision Geometric Camera Calibration.
Numerical Analysis – Eigenvalue and Eigenvector Hanyang University Jong-Il Park.
*Man-Cheol Kim, Hyung-Jo Jung and In-Won Lee *Man-Cheol Kim, Hyung-Jo Jung and In-Won Lee Structural Dynamics & Vibration Control Lab. Structural Dynamics.
2.1 – Linear and Quadratic Equations Linear Equations.
Large-Scale Matrix Factorization with Missing Data under Additional Constraints Kaushik Mitra University of Maryland, College Park, MD Sameer Sheoreyy.
1 Chapter 7 Numerical Methods for the Solution of Systems of Equations.
STATIC ANALYSIS OF UNCERTAIN STRUCTURES USING INTERVAL EIGENVALUE DECOMPOSITION Mehdi Modares Tufts University Robert L. Mullen Case Western Reserve University.
Toward an Automatically Tuned Dense Symmetric Eigensolver for Shared Memory Machines Yusaku Yamamoto Dept. of Computational Science & Engineering Nagoya.
V.M. Sliusar, V.I. Zhdanov Astronomical Observatory, Taras Shevchenko National University of Kyiv Observatorna str., 3, Kiev Ukraine
F. Fairag, H Tawfiq and M. Al-Shahrani Department of Math & Stat Department of Mathematics and Statistics, KFUPM. Nov 6, 2013 Preconditioning Technique.
Performance of BLAS-3 Based Tridiagonalization Algorithms on Modern SMP Machines Yusaku Yamamoto Dept. of Computational Science & Engineering Nagoya University.
Circuit Simulation using Matrix Exponential Method Shih-Hung Weng, Quan Chen and Chung-Kuan Cheng CSE Department, UC San Diego, CA Contact:
Finding Rightmost Eigenvalues of Large, Sparse, Nonsymmetric Parameterized Eigenvalue Problems Minghao Wu AMSC Program Advisor: Dr. Howard.
ALGEBRAIC EIGEN VALUE PROBLEMS
Geometric Camera Calibration
Multiplicative updates for L1-regularized regression
Solving Systems of Linear Equations: Iterative Methods
Eigenspectrum calculation of the non-Hermitian O(a)-improved Wilson-Dirac operator using the Sakurai-Sugiura method H. Sunoa, Y. Nakamuraa,b, K.-I. Ishikawac,
Chapter 3 Component Reliability Analysis of Structures.
FTCS Explicit Finite Difference Method for Evaluating European Options
Parallel Inversion of Polynomial Matrices
Chapter 27.
Results (Accuracy of Low Rank Approximation by iSVD)
한국지진공학회 춘계 학술발표회 서울대학교 호암교수관 2003년 3월 14일
Numerical Analysis Lecture11.
ECE 576 POWER SYSTEM DYNAMICS AND STABILITY
Presentation transcript:

A Large-Grained Parallel Algorithm for Nonlinear Eigenvalue Problems Using Complex Contour Integration Takeshi Amako, Yusaku Yamamoto and Shao-Liang Zhang Dept. of Computational Science & Engineering Nagoya University, Japan

Outline of the talk Introduction The nonlinear eigenvalue problem Existing algorithms Our objective The algorithm Formulation as a nonlinear equation Application of Kravanja et al s method Detecting and removing spurious eigenvalues Numerical results Accuracy of the computed eigenvalues Parallel performance Conclusion

Introduction The nonlinear eigenvalue problem Given A(z) C n × n, z: complex parameter Find z 1 C such that A(z 1 ) x = 0 has a nonzero solution x = x 1. z 1 and x 1 are called the eigenvalue and the corresponding eigenvector, respectively. Examples A(z) = A – zB + z 2 C : quadratic eigenvalue problem A(z) = A – zB + e z C : general nonlinear eigenvalue problem Applications Electronic structure calculation Nonlinear elasticity Theoretical fluid dynamics

Existing algorithms Multivariate Newtons method and its variants Locally quadratic convergence Requires good initial estimate both for z 1 and x 1. Nonlinear Arnoldi methods Nonlinear Jacobi-Davidson methods Efficient for large sparse matrices Not suitable for finding all eigenvalues within a specified region of the complex plane difficult to obtain

Our objective Let : closed Jordan curve on the complex plane, A(z) C n × n : analytical function of z in We propose an algorithm that can find all the eigenvalues within, and has large-grain parallelism. Im z Re z O Assumption: In the following, we mainly consider the case where is a circle centered at the origin and with radius r. Related work: Sakurai et al. propose an algorithm for linear generalized eigenvalue problems r

Our approach The basic idea Let f(z) = det(A(z)). Then f(z) is an analytical function of z in and the eigenvalues of A(z) are characterized as the zeros of f(z). Use Kravanjas method (Kravanja et al., 1999) to find the zeros of an analytic function.

Finding zeros of f(z) Let z 1, z 2,..., z m : zeros of f(z) in, and 1, 2,..., m : their multiplicity. Then f(z) can be written as Define the complex moments by Then f(z) = ×g(z) analytical and nonzero in analytical in unknowncomputable

Finding zeros of f(z) (cont'd) To extract information on {z k } from { p }, define the following matrices: Then it is easy to see that

Finding zeros of f(z) (cont'd) Noting that V m and D m are nonsingular, we have the following equivalence relation: That is, we can find the zeros of f(z) in by computing the complex moments 0, 1,..., 2m-1, constructing H m and H m <, and computing the eigenvalues of H m < – H m. is an eigenvalue of H m < – H m is an eigenvalue of m – k, = z k

Application to the nonlinear eigenvalue problem In our case, f(z) = det(A(z)) and By applying the trapezoidal rule with K points, we have where Im z Re z O

The algorithm The computationally intensive part. Large-grain parallelism

Detecting and removing spurious eigenvalues Usually, we do not know m, the number of eigenvalues of A(z) in, in advance and use some estimate M instead. When M > m, the eigenvalues of H m < – H m include spurious solutions that do not correspond to an eigenvalue of A(z). To detect them, we compute the corresponding eigenvector by inverse iteration and evaluate the relative residual defined by Of course, this quantity can also be used to check the accuracy of the computed eigenvalues. relative residual =

Numerical results Test problem A(z) = A – zI + B(z), where A(z) : real random nonsymmetric matrix B(z) : antidiagonal matrix with antidiagonal elements e z : parameter to specify the strength of nonlinearity Parameters n = 500, 1000, , 10 –4, 10 –3, 10 –2, 10 –1 Computational environment Fujitsu HPC2500 (SPARC 64IV), 1-16 processors Program written with C and MPI LAPACK routines were used to compute (A(z)) –1 and to compute the eigenvalues of H m < – H m.

Accuracy of the computed eigenvalues Parameters n = 500 and = 0.1 r = 0.85, K = 128 and M = 11. There are 7 eigenvalues in. Results Our algorithm succeeded in locating all the eigenvalues in. The relative residuals were all under 10 –10. Similar results for other cases. Im z Re z

Effect of K and M on the accuracy Effect of the number of sample points K Usually K=128 gives sufficient accuracy. Effect of the Hankel matrix size M It is better to take M a few more than the number of eigenvalues within (7 in this case). This is to mitigate the perturbation from eigenvalues outside. Residuals as a function of K. Residuals as a function of M. KM

Detecting and removing spurious eigenvalues Parameters n = 1000 and = 0.01 r = 0.7, K = 128 and M = 10. There are 9 eigenvalues in. Eigenvalues of H m < – H m 10 eigenvalues were found within. For 9 of the eigenvalues, the residual was less than 10 –11. For one eigenvalue, the residual was 10 –2. Im z Re z spurious eigenvalue

Parallel performance Performance on Fujitsu HPC2500 Matrix size: n = 500, 1000, 2000 Number of processors: P = 1, 2, 4, 8, 16 Execution time (sec) Number of processors Almost linear speedup was obtained in all cases due to large-grain parallelism.

Parallel performance (cont'd) Performance in a Grid environment Matrix size: n = 1000 Machine: Intel Xeon Cluster Master-worker type parallelization using OmniRPC (GridRPC) Execution time Number of processors Good scalability was obtained for up to 14 processors. 0:00:00 0:30:00 1:00:00 1:30:00 2:00: Speedup

Summary of this study We proposed a new algorithm for the nonlinear eigenvalue problem based on complex contour integration. Our algorithm can find all the eigenvalues within a closed curve on the complex plane. Moreover, it has large-grain parallelism and is expected to show excellent parallel performance. These advantages have been confirmed by numerical experiments.

Future work Performance evaluation on large-scale grid environments. Application to practical problems. Computation of scaling exponent in theoretical fluid dynamics Development of an efficient algorithm for computing