Parallel Direct Methods for Sparse Linear Systems

Slides:



Advertisements
Similar presentations
E. T. S. I. Caminos, Canales y Puertos1 Engineering Computation Lecture 6.
Advertisements

Chapter 2 Simultaneous Linear Equations
CSCI-455/552 Introduction to High Performance Computing Lecture 25.
Linear Systems LU Factorization CSE 541 Roger Crawfis.
MATH 685/ CSI 700/ OR 682 Lecture Notes
Linear Systems of Equations
Solving Linear Systems (Numerical Recipes, Chap 2)
Lecture 7 Intersection of Hyperplanes and Matrix Inverse Shang-Hua Teng.
LU Factorization LU-factorization Matrix factorization Forward substitution Back substitution.
Systems of Linear Equations
SOLVING SYSTEMS OF LINEAR EQUATIONS. Overview A matrix consists of a rectangular array of elements represented by a single symbol (example: [A]). An individual.
Solution of linear system of equations
1cs542g-term Notes  Assignment 1 is out (questions?)
Linear Systems of Equations Ax = b Marco Lattuada Swiss Federal Institute of Technology - ETH Institut für Chemie und Bioingenieurwissenschaften ETH Hönggerberg/
Sparse Matrix Methods Day 1: Overview Matlab and examples Data structures Ax=b Sparse matrices and graphs Fill-reducing matrix permutations Matching and.
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display. Parallel Programming in C with MPI and OpenMP Michael J. Quinn.
Monica Garika Chandana Guduru. METHODS TO SOLVE LINEAR SYSTEMS Direct methods Gaussian elimination method LU method for factorization Simplex method of.
資訊科學數學11 : Linear Equation and Matrices
Elementary Linear Algebra Howard Anton Copyright © 2010 by John Wiley & Sons, Inc. All rights reserved. Chapter 1.
Elementary Linear Algebra Howard Anton Copyright © 2010 by John Wiley & Sons, Inc. All rights reserved. Chapter 1.
Elementary Linear Algebra
Scientific Computing Linear Systems – LU Factorization.
Eigenvalue Problems Solving linear systems Ax = b is one part of numerical linear algebra, and involves manipulating the rows of a matrix. The second main.
Copyright © 2011 Pearson, Inc. 7.3 Multivariate Linear Systems and Row Operations.
MA2213 Lecture 5 Linear Equations (Direct Solvers)
Linear Systems Gaussian Elimination CSE 541 Roger Crawfis.
MATH 250 Linear Equations and Matrices
Sec 3.2 Matrices and Gaussian Elemination Coefficient Matrix 3 x 3 Coefficient Matrix 3 x 3 Augmented Coefficient Matrix 3 x 4 Augmented Coefficient Matrix.
Chapter 3 Solution of Algebraic Equations 1 ChE 401: Computational Techniques for Chemical Engineers Fall 2009/2010 DRAFT SLIDES.
Lecture 8 Matrix Inverse and LU Decomposition
Triangular Form and Gaussian Elimination Boldly on to Sec. 7.3a… HW: p odd.
Linear algebra: matrix Eigen-value Problems Eng. Hassan S. Migdadi Part 1.
Chapter 5 MATRIX ALGEBRA: DETEMINANT, REVERSE, EIGENVALUES.
Solving Linear Systems Solving linear systems Ax = b is one part of numerical linear algebra, and involves manipulating the rows of a matrix. Solving linear.
Solution of Sparse Linear Systems
Matrices and Systems of Equations
Meeting 19 System of Linear Equations. Linear Equations A solution of a linear equation in n variables is a sequence of n real numbers s 1, s 2,..., s.
ECE 530 – Analysis Techniques for Large-Scale Electrical Systems Prof. Hao Zhu Dept. of Electrical and Computer Engineering University of Illinois at Urbana-Champaign.
1 Chapter 7 Numerical Methods for the Solution of Systems of Equations.
Linear Systems Dinesh A.
Programming Massively Parallel Graphics Multiprocessors using CUDA Final Project Amirhassan Asgari Kamiabad
 Recall that when you wanted to solve a system of equations, you used to use two different methods.  Substitution Method  Addition Method.
Gaoal of Chapter 2 To develop direct or iterative methods to solve linear systems Useful Words upper/lower triangular; back/forward substitution; coefficient;
2 2.5 © 2016 Pearson Education, Ltd. Matrix Algebra MATRIX FACTORIZATIONS.
Linear Algebra Engineering Mathematics-I. Linear Systems in Two Unknowns Engineering Mathematics-I.
Matrices, Vectors, Determinants.
Numerical Computation Lecture 6: Linear Systems – part II United International College.
Algebra 3 5.1/2 Systems of Linear Equations/Matrices.
Multivariable Linear Systems and Row Operations
Systems of linear equations
Lecture 2 Matrices Lat Time - Course Overview
Spring Dr. Jehad Al Dallal
5 Systems of Linear Equations and Matrices
CS 290N / 219: Sparse Matrix Algorithms
Linear Equations.
Linear Algebra Lecture 15.
Math Linear Algebra Introduction
Instruction of Chapter 8 on
Chapter 10: Solving Linear Systems of Equations
Chapter 1 Systems of Linear Equations and Matrices
CSCE569 Parallel Computing
Review of Matrix Algebra
Numerical Computation and Optimization
Numerical Analysis Lecture10.
6 Systems of Linear Equations and Matrices
Linear Systems Numerical Methods.
Lecture 8 Matrix Inverse and LU Decomposition
Ax = b Methods for Solution of the System of Equations:
Outline Sparse Reconstruction RIP Condition
Ax = b Methods for Solution of the System of Equations (ReCap):
Presentation transcript:

Parallel Direct Methods for Sparse Linear Systems

Contents Problem Statement Motivation Types of Algorithms Sparse Matrices Methods to solve Sparse Matrices

Problem Statement

Problem Statement The solution of the linear system is the values of the unknown vector x, that satisfy every equation in the linear system for the given matrix A and vector b

Why Linear Systems? How many liters of a 70% alcohol solution must be added to 50 liters of a 40% alcohol solution to produce a 50% alcohol solution? Manufacturing companies use a system of linear equations to solve these kind of problems in their production line. Oil and Gas companies makes adjustments to their recipes for manufacturing gas and diesel based on the component prices. Even hotdog companies do that and so on… All these problems can be solved using linear system solvers.

Algorithms There has been several algorithms proposed to solve linear systems. They can be categorized as Direct Methods Gaussian Algorithm – O(n3) Cholesky Factorization etc., Iterative Methods Jacobi Method The Conjugate Gradient Method

The Gauss method –Multiplying an equation by a nonzero constant, The main idea of the method is by using the equivalent operations to transform a dense system into an upper triangular system, which can be easily solved Equivalent transformations: –Multiplying an equation by a nonzero constant, –Interchanging two equations, –Adding an equation multiplied by a constant to another equation

The Gauss method On the first stage of the algorithm, which is called the Gaussian elimination, the initial system of linear equations is transformed into an upper triangular system by the sequential elimination of unknowns: Ux = c On the second stage of the algorithm, which is called the back substitution, the values of the variables are calculated

Example Complexity is O(n3) Some real time applications have massive sizes for A. It will take years to solve such linear systems using Gaussian Method. This method is not well-suited for sparse systems. Increases storage and total Flops count.

Sparse System A system of linear equations is called sparse if only a relatively small number of its matrix elements aij are nonzero. It is wasteful to use general methods of linear algebra on such problems, because most of the O(N3) arithmetic operations devoted to solving the set of equations or inverting the matrix involve zero operands.

Sparse matrices arise in ... computational fluid dynamics, finite-element methods, statistics, time/frequency domain circuit simulation, dynamic and static modeling of chemical processes, cryptography, magneto-hydrodynamics, electrical power systems, differential equations, quantum mechanics, structural mechanics (buildings, ships, aircraft, human body parts...), heat transfer, MRI reconstructions, vibroacoustics, linear and non-linear optimization, financial portfolios, semiconductor process simulation, economic modeling, oil reservoir modeling, astrophysics, crack propagation, Google page rank, 3D computer vision, cell phone tower placement, tomography, multibody simulation, model reduction, nano-technology, acoustic radiation, density functional theory, quadratic assignment, elastic properties of crystals, natural language processing, DNA electrophoresis, ...

Example in structural mechanics:

Sparse data structures compressed sparse column format

Factorization process Solution of Ax = b A is unsymmetric : A is factorized as: A = LU, where L is a lower triangular matrix, and U is an upper triangular matrix. Forward-backward substitution: Ly = b then Ux = y A is symmetric: positive definite A = LLT

Cholesky Factorization Algorithm Cholesky Factorization is a variant of Gaussian elimination that takes advantage of symmetry to reduce both work and storage by about half.

Next Presentation…. Discuss about “SuiteSparse”, an API which has implementations of various algorithms for solving Sparse Systems. How GPU acceleration can be achieved for Cholesky Factorization? Complexity analysis of Sequential and Parallel versions of the algorithm.

QUESTIONS ?