matlab优化

Pascal Getreuer, January 2008

Contents

1 The Profiler 2

2 Array Preallocation 4

3 JIT Acceleration 6

4 Vectorization 8

4.1 Vectorized Computations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

4.2 Vectorized Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

5 Inlining Simple Functions 13

9.1 Clip a value without using if statements . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

9.2 Convert any array into a column vector . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

9.3 Find the min/max of a matrix or N-d array . . . . . . . . . . . . . . . . . . . . . . . . . 25

9.4 Flood filling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

9.5 Vectorized use of set on GUI objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

10 Further Reading 27

Introduction

Matlab code is interpreted during runtime. Languages like C and Fortran are faster because they are

first compiled into the computer’s native language. Interpreting in realtime enables better platform

independence, greater language flexibility, and interactive debugging, but at the cost of slower speed,

Caution!

Learn the language first. Optimization requires comfort with the syntax and features of the

language. This article is not a tutorial on MATLAB.

Use comments. Optimized code tends to be terse and cryptic. Help others and yourself by

remembering to comment.

Don’t optimize code before its time. Is optimization worth the effort? If the code will soon

be revised or extended, it will be rewritten anyway.

Only optimize where necessary. Focus your efforts on bottlenecks.

1

1

The Profiler

For example, consider profiling the following function:

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

function result = example1(Count)

for k = 1:Count

end

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

run example1 once before profiling. The profiler report command shows a report. Depending on

the system, profiler results may differ from this example.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

MATLAB Profile Report: Summary

Function List

2

The most time-consuming lines are displayed, along with time, time percentage, and line number. The

most costly lines are the computations on lines 4 and 7.

Another helpful section of the profile report is “M-Lint Results,” which gives analysis from the M-Lint

code analyzer. Potential problems and suggested improvements are listed here.

M-Lint results

times with a loop:

Adjust the number of loop iterations so that the time it takes to run is noticeable. More iterations

yields better time resolution.

3

2

Array Preallocation

Matlab’s matrix variables have the ability to dynamically augment rows and columns. For example,

a =

Consider the code:

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

a(1) = 1;

b(1) = 0;

for k = 2:8000

*

b(k − 1);

*

b(k − 1);

end

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

b(1) = 0;

for k = 2:8000

*

b(k − 1);

*

b(k − 1);

end

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

With this modification, the code takes only 0.14 seconds (over three times faster). Preallocation is often

easy to do, in this case it was only necessary to determine the right preallocation size and add two lines.

4