Parallel and Domain Decomposed Marching Method for Poisson Equation 大氣五 b94209002 簡睦樺.

Slides:



Advertisements
Similar presentations
The Maximum Likelihood Method
Advertisements

Fixed Points and The Fixed Point Algorithm. Fixed Points A fixed point for a function f(x) is a value x 0 in the domain of the function such that f(x.
Rules of Matrix Arithmetic
Partial Differential Equations
CSE 330: Numerical Methods
Chapter 8 Elliptic Equation.
Algorithms Recurrences. Definition – a recurrence is an equation or inequality that describes a function in terms of its value on smaller inputs Example.
1.5 Elementary Matrices and a Method for Finding
1 The RSA Algorithm Supplementary Notes Prepared by Raymond Wong Presented by Raymond Wong.
Infinite Sequences and Series
1cs542g-term Notes  Simpler right-looking derivation (sorry):
Curve-Fitting Regression
Approximation on Finite Elements Bruce A. Finlayson Rehnberg Professor of Chemical Engineering.
Preference Analysis Joachim Giesen and Eva Schuberth May 24, 2006.
Chi Square Distribution (c2) and Least Squares Fitting
Partial differential equations Function depends on two or more independent variables This is a very simple one - there are many more complicated ones.
Linear Simultaneous Equations
CISE-301: Numerical Methods Topic 1: Introduction to Numerical Methods and Taylor Series Lectures 1-4: KFUPM.
EXAMPLE 2 Solve a matrix equation SOLUTION Begin by finding the inverse of A = Solve the matrix equation AX = B for the 2 × 2 matrix X. 2 –7 –1.
Non-Linear Simultaneous Equations
Using Inverse Matrices Solving Systems. You can use the inverse of the coefficient matrix to find the solution. 3x + 2y = 7 4x - 5y = 11 Solve the system.
1 1.1 © 2012 Pearson Education, Inc. Linear Equations in Linear Algebra SYSTEMS OF LINEAR EQUATIONS.
Functions. A function is a relation that has exactly one output for each input.
Solving Systems of Linear Equations and Inequalities
3 DERIVATIVES.
Physics 114: Lecture 15 Probability Tests & Linear Fitting Dale E. Gary NJIT Physics Department.
Solving Systems of Linear Equations by Graphing
I can solve systems of equations by graphing and analyze special systems.
880.P20 Winter 2006 Richard Kass 1 Confidence Intervals and Upper Limits Confidence intervals (CI) are related to confidence limits (CL). To calculate.
Review Solve each equation or inequality b = 2, -5 Multiply every term by 12.
 The functions that we have met so far can be described by expressing one variable explicitly in terms of another variable.  For example,, or y = x sin.
A.Broumandnia, 1 3 Parallel Algorithm Complexity Review algorithm complexity and various complexity classes: Introduce the notions.
3 DIFFERENTIATION RULES. The functions that we have met so far can be described by expressing one variable explicitly in terms of another variable. 
5-3 Equations as Relations
Scientific Computing Partial Differential Equations Implicit Solution of Heat Equation.
Number Theory Project The Interpretation of the definition Andre (JianYou) Wang Joint with JingYi Xue.
CY3A2 System identification
Boundary Value Problems l Up to this point we have solved differential equations that have all of their initial conditions specified. l There is another.
4.1 Antiderivatives and Indefinite Integration Definition of Antiderivative: A function F is called an antiderivative of the function f if for every x.
12 INFINITE SEQUENCES AND SERIES. In general, it is difficult to find the exact sum of a series.  We were able to accomplish this for geometric series.
This screen shows two lines which have exactly one point in common. The common point when substituted into the equation of each line makes that equation.
1 EEE 431 Computational Methods in Electrodynamics Lecture 18 By Dr. Rasime Uyguroglu
Copyright © 2006 The McGraw-Hill Companies, Inc. Permission required for reproduction or display. by Lale Yurttas, Texas A&M University Chapter 271 Boundary-Value.
Divide and Conquer. Recall Divide the problem into a number of sub-problems that are smaller instances of the same problem. Conquer the sub-problems by.
Copyright © Cengage Learning. All rights reserved.
Module 7 Test Review. Formulas Shifting Functions  Given g(x) on a graph, f(x) = g(x) + 7 would mean that…..  Answer: g(x) would be shifted 7 units.
Marching Solver for Poisson Equation 大氣四 簡睦樺. Outline A brief review for Poisson equation and marching method Parallel algorithm and consideration for.
Linear Systems Numerical Methods. 2 Jacobi Iterative Method Choose an initial guess (i.e. all zeros) and Iterate until the equality is satisfied. No guarantee.
Objective: To apply the Law of Cosines for finding the length of a missing side of a triangle. Lesson 18 Law of Cosines.
1 Spring 2003 Prof. Tim Warburton MA557/MA578/CS557 Lecture 25.
Richard Kass/F02P416 Lecture 6 1 Lecture 6 Chi Square Distribution (  2 ) and Least Squares Fitting Chi Square Distribution (  2 ) (See Taylor Ch 8,
Modified Bessel Equations 홍성민. General Bessel Equation of order n: (1) The general solution of Eq.(1) Let’s consider the solutions of the 4.
Chapter 3 Systems of Equations. Solving Systems of Linear Equations by Graphing.
CSE 330: Numerical Methods. What is true error? True error is the difference between the true value (also called the exact value) and the approximate.
Copyright © Cengage Learning. All rights reserved The Integral Test and Estimates of Sums.
EEE 431 Computational Methods in Electrodynamics
We will be looking for a solution to the system of linear differential equations with constant coefficients.
Solving Systems of Linear Equations and Inequalities
Copyright © Cengage Learning. All rights reserved.
Tables and Relations Review
Equations as Relations
Copyright © Cengage Learning. All rights reserved.
MATH 1310 Session 2.
Copyright © Cengage Learning. All rights reserved.
Solve the equation: 6 x - 2 = 7 x + 7 Select the correct answer.
Tables and Relations Review
Solving for x and y when you have two equations
RELATIONS & FUNCTIONS CHAPTER 4.
Programming assignment #1 Solving an elliptic PDE using finite differences Numerical Methods for PDEs Spring 2007 Jim E. Jones.
Presentation transcript:

Parallel and Domain Decomposed Marching Method for Poisson Equation 大氣五 b 簡睦樺

Outline Basic Marching Method for elliptic problem. Domain Decomposed Algorithm Analysis and Operation Counts for this Algorithm Results

Model Problem A given open domain of R n satisfies the following equation for n=1,2,3, f and g are given, and

Finite difference approach With above approach, we can rewrite equation Note that derivative of x and z also use above approach. For i,j,k=0,1,…,N+1, i,j,k=0, N+1 are the boundary condition given.

Approach Idea If we have the exact solution u and plug into the approach formula, then it should satisfy the equation but there will emerge error with O(δx 2 ). On the other words, if there exist a data satisfies the plug into the formula, then there must have error at each point.

Testing Result Whether these error comes from a basis? The answer is true.

How to solve these error? If there exist enough exact data, then the behavior must be control. That is to say, we use the boundary data and have an initial guess next to the boundary points; thus, the error should only exist at the point which comes form the formula marching.

Write down the formula of error u i,j is the exact solution, e is the error.

Error system We can write the system of error, since they march independently. Collect each error result at the end of point, we can correct the error by boundary condition of another side. Solve this system, we can get the initial error and correct the marching data by exact solution next to boundary.

Domain Decomposed Algorithm Although the above method is prefect in idea case, the condition number of the system may easily blow up. We need another method to solve the problem.

Idea of Domain Decomposed We can find that if we have the marching value at same point, then it must have same value.

The relationship between two initial value

System of Error

Analysis for Algorithm Basic Algorithm 1.Data march all domain O(NX*NY*NZ) 2.Error march all domain O(NX^2*NZ^2*NY) 3.Compute difference of boundary value O(NX*NZ) 4.Solve the system O((NX*NZ)^3) 5.Data march again O(NX*NY*NZ)

An efficient Idea Solve the system, and save the inverse matrix. We need only to take the matrix product when need it. The following is the operation count: 1. Data march all domain O(NX*NY*NZ) 2. Take matrix product O((NX*NZ)^2) 3. Data march again O(NX*NY*NZ) The algorithm has a upper bound O(N^4).

Domain Decomposed Case Unfortunately, domain decomposed case require to solve a larger system than before. Although we prove that the domain decomposed case has a same upper bound compared to usually problem, but the lower bound is rather small than before. i.e.

Analysis 1.Data march all domain O(NX*NY*NZ) 2.Error march all domain O(NX^2*NZ^2*NY) 3.Compute difference of boundary value O(NX*NZ*NY) 4.Solve the system O((NX*NY*NZ)^3) 5.Data march again O(NX*NY*NZ) This step is relative to size of sub-domain.

The efficient solver 1. Data march all domain O(NX*NY*NZ) 2. Take matrix product O((NX*NY*NZ)^2) 3. Data march again O(NX*NY*NZ) The algorithm has a upper bound O(N^6). Remark. NY>>NB

Result for 3D equation

Convergent table and log-log plot Mesh sizeErrorRatioObserved order e-3NaN e e e Least squares fit gives E(h) = * h^1.54

24*24*24 Case NodesTime(Total)Step1Step2Step Least squares fit gives E(h) = 18.0 * h^0.8

Thank you