Fast Multipole Methods: Fundamentals and Applications

 

Instructors: Ramani Duraiswami  & Nail A. Gumerov

Course No. CMSC 878R/AMSC 698R/MAIT  627

E-mail: ramani AT umiacs.umd.edu;

gumerov AT umiacs.umd.edu

Spring  2008, Tuesdays and Thursdays, 3:30 p.m. – 4:45 p.m.,

Location: CSI 3118

 

Homework is given on Wednesdays at the end of the class and due on the next Wednesday at the beginning of class

Lecture Notes from previous years 2006 2004 2003 2002

(Some of the material here is now also described in the book  Fast Multipole Methods for the Helmholtz Equation in Three Dimensions (The Elsevier Electromagnetism Series))

This course will involve a project – for maximum benefit you should start looking for a project early in the course.

The course outline below is a tentative organization of the material. It is subject to change!

DATE

LECTURE

CONTENTS

READING

01/29/2008

 

Lecture1

Introduction, Course rules, exams

Structure and Efficiency

Algorithm Complexity and the solution of Large Problems

Applications: Physics, Computer Vision, etc.

·        G. Moore, "Cramming More Components onto integrated circuits," Electronics, 38, 1965

·       Steele Prize Announcements, 2001

·        Dongarra and Sullivan, 2000

01/31/2008

Lecture 2

Simple example of factorization for degenerate kernel

Asymptotic Complexity. Examples of complexity

Structured Matrices, Fast Fourier Transform

C. Moler (Chapter on the FFT)

Homework 1

(due 02/07/2008)

Solution

Simple factorization example with a degenerate kernel.

Remember to label axes in all plots.

 

02/05/2008

Lecture 3

Factorization of non degenerate Kernels. Local and Far-field expansions. Examples. Taylor Series.

D. Donoho, "High-Dimensional Data Analysis: The Curses and Blessings of Dimensionality," AMS Lecture at a Conference on "Math Challenges of the 21st Century"

02/07/2008

Lecture 4

 

 

 

 

Homework 2

 

Solution

Simple factorization example with a nondegenerate kernel and truncated expansions with error bounds.

Remember to label axes in all plots.

 

02/12/2008

Lecture 5

Complexity of the “middleman” approach considering errors. Grouping and exclusion to deal with non-factorizable points. “Pre-FMM” algorithm with O(N3/2) complexity.

 

02/14/2008

Lecture 6

 

 

Example of the (x-y)-1 kernel. Translation operators.

 

Homework 3

(due 02/21/2008)

 

Pre-FMM … Introduces the grouping and factorization ideas. Simple data structure for identifying points in a 1-D “box” via sorting

 

02/19/2008

Lecture 7

 

 

02/21/2008

Lecture 8

 

Homework 4

 

 

02/26/2008

Lecture 9

Demo based on code written by Yang Wang and Trung Nguyen as projects for the course

Java FMM animation

You may need to download java from java

02/28/2008

Lecture 10

 

Homework 5

Complexity of the MLFMM

 

03/04/2008

 

No class: Attend the following talk in MTH 3206

Professor Xiaobai Sun: Efficient Encoding of Irregular Geometric Structures in Matrix Computations

Reading Material: X. Sun and N. Pitsianis: Matrix Version of the Fast Multipole Method, SIAM Review, pp. 289-300, 2001w

03/06/2008

Lecture 11

 

 

Data Structures for the FMM

 

03/11/2008

Lecture 12

(Click on previous lecture)

Data Structures for the FMM

 

 

Homework 6

Data Structures

 

03/13/2008

Lecture 13

Error bounds for the FMM

 

03/18/2008

03/20/2008

Spring Break

No Class

 

03/25/2008

Lecture 14

Iterative Methods Lecture 1:

 

03/27/2008

Lecture 15

 

Homework 7

 

 

04/01/2008

Lecture 16

Laplace equation

Projects

04/03/2008

Lecture 17

 

Proj Report 1

Faster translation methods

 

04/08/2008

Lecture 18

IFGT

 

04/10/2008

Lecture 19

IFGT + BEM

 

04/15/2008

Lecture 20

 

Lecture20a

 

 

Proj Report 2

Adaptive FMM

+ discussion of a second adaptive algorithm

Animations

04/17/2008

Lecture 21

 

 

04/22/2008

Lecture 22

 

 

04/24/2008

Lecture 23

 

 

04/29/2008

Lecture 24

 

 

05/01/2008

Lecture 25

Helmholtz equation

 

05/03/2008

Lecture 26

Some issues with the FMM

 

05/08/2008

Lecture 27

 Non uniform FFT

 

FFTM

 

05/10/2008

Lecture 28

Project Presentations

 

 

 

 

 

05/21/2008

Wednesday

Final Exam

10:30am-12:30 pm (according to Testudo)

Project Reports due

 

 

Other References

 

L. Greengard and J. Strain, "The Fast Gauss Transform," SIAM J. Sci. and Stat. Comput. 12, 79 (1991)