parallel and domain decomposed marching method for poisson equation

22
Parallel and Domain Decomposed Marching Method for Poisson Equation 大大大 b94209002 大大大

Upload: hayley

Post on 06-Jan-2016

16 views

Category:

Documents


0 download

DESCRIPTION

Parallel and Domain Decomposed Marching Method for Poisson Equation. 大氣五 b94209002 簡睦樺. Outline. Basic Marching Method for elliptic problem. Domain Decomposed Algorithm Analysis and Operation Counts for this Algorithm Results. Model Problem. - PowerPoint PPT Presentation

TRANSCRIPT

Page 1: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Parallel and Domain Decomposed Marching Method

for Poisson Equation大氣五 b94209002 簡睦樺

Page 2: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Outline

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

Algorithm• Results

Page 3: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Model Problem

• A given open domain of Rn satisfies the following equation

for n=1,2,3, f and g are given, and

ugu

openufu n

,

,

2

2

ix

Page 4: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Finite difference approach

22

11

2

2 ,,,,2,,,,xO

x

zyxuzyxuzyxu

x

zyxu kjikjikjikji

With above approach, we can rewrite equation

2

,,2

2

,,2

,1,,,2

,1, 2z

u

x

uuufyu kjikji

kjikjiijkkji

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.

Page 5: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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(δx2).

• On the other words, if there exist a data satisfies the plug into the formula, then there must have error at each point.

Page 6: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Testing Result

Whether these error comes from a basis?

The answer is true.

Page 7: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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.

Page 8: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Write down the formula of error

2

,2

1,,1,

,,,

2

ˆ

x

eeee

euu

jijijiji

jijiji

u i,j is the exact solution, e is the error.

Page 9: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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.

Page 10: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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.

Page 11: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Idea of Domain Decomposed

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

Page 12: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

The relationship between two initial value

rrll eueu ˆˆ

Page 13: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

System of Error

Page 14: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Analysis for AlgorithmBasic 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)

Page 15: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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).

Page 16: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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.

Page 17: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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.

Page 18: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

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

Page 19: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Result for 3D equation

Page 20: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Convergent table and log-log plotMesh size Error Ratio Observed order

0.5 1.062e-3 NaN NaN

0.25 4.685e-4 2.27 1.18

0.125 1.532e-4 3.06 1.61

0.0625 4.385e-5 3.49 1.80

Least squares fit gives E(h) = 0.003 * h^1.54

Page 21: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

24*24*24 CaseNodes Time(Total) Step1 Step2 Step3

1 49.78 0.17 49.32 0.22

2 33.25 0.08 32 0.18

4 17.25 0.04 17 0.9

Least squares fit gives E(h) = 18.0 * h^0.8

Page 22: Parallel and Domain Decomposed  Marching Method  for Poisson Equation

Thank you