Octave tutorial

Juan Carlos Chaves

Informations

• Publié par: Mezat
• Nombre de lectures: 171
• Langue: English

Extrait

Non-traditional High-Level Languages for HPC Applications:

GNU Octave

Juan Carlos Chaves, Ph.D. SIP On-site at ARL MSRC Ohio Supercomputer Center

Users Group Conference (UGC2006) 26 Jun 2006

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

1

Purpose of This Tutorial

• Provide an introduction to the Octave system

• Teach how to obtain and configure Octave for your own PC

• Teach how to configure, build and install Octave at the HPCMP HPC platforms

• Teach how to exploit Octave HPC technology at the HPCMP Centers

• Be a concise guide to relevant Octave info on the Web

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

2

•

•

•

•

What is Octave?

High-level language, primarily intended for numerical computations initiated and maintained byJohn W. Eaton(University of Wisconsin)

A stated goal of Octave isto be compatible with MATLAB

Provides aconvenient command line interactive interfacefor solving scientific and technical problems numerically

Has built-ina large collection of mathematical algorithms and convenience functions

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

3

•

•

•

Typical uses of Octave

Prototyping:an interactive Octave session becomes a system-prototyping environment rivaling systems such as the well-known MATLAB

Programming language:a powerful but easy to use programming language, mostly compatible with MATLAB, for use in developing sophisticated programs and algorithms

Batch-oriented HPC applications:Octave and its parallel extensions may also be run under a batch system such as LSF, the mandate queuing system at the DoD HPCMP Centers

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

4

•

•

•

•

Octave Advantages

Free programavailable under the GNU General Public License (GNU GPL) and runs on Windows, Linux and Mac OS Xincluding architectures unsupported by MATLAB

MATLAB clone:syntax is very similar to MATLAB. Careful programming will allow a script to run on both Octave and MATLAB

Open source community: large and very active support communityall over the world

High productivity system:interactive - code development proceeds incrementally. Excellent development and rapid prototyping environment

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

5

•

•

•

•

•

Octave Disadvantages

Relativelyomc ’ ns iexplioatlltadecorp neru(But fortunately, you re taking this tutorial!)

Unlike MATLAB, it does not support object-oriented programming

Unlike MATLAB supportsonly basic graphic capabilities, limiting the application of Octave in scientific and engineering graphics, including the creation of GUIs.

Unlike MATLAB, lacks support from a multi-million dollar companysuch as the MathWorks, Inc.

Unlike MATLAB,fewer toolboxes are available

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

6

Octave – basic features

•No declarations needed: octave:1> 14 + 28 ans = 42

•Mixed types: octave:2> product = 16*16.34 product = 261.44

•Semicolons suppress output: octave:5> product = 16*16.99; octave:6> product product 271.84 =

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

7

Octave – Matrix notation

•Working with matrices: octave:7> a = [ 1, 1, 2; 3, 5, 8; 13, 21, 34 ] a = 1 1 2 3 5 8 13 21 34

•Row vectors and the colon operator : _ octave:8> row vector = 1:5 row vector = _

1 2 3 4 5

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

8

Octave – Matrix Subscripting

•Working with subscripts: octave:13> b = rand (3, 3); octave:14> b(3,1), b(3,2), b(3,3) ans = 0.27638 ans = 0.94927 ans = 0.36208

octave:15> b(3,2) = 1;   octave:16> b b = 0.46079 0.31457 0.53600 0.55636 0.22555 0.11940 0.276381.000000.36208

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

9

Octave – Multidimensional Arrays • Example of a 3D random array filled with normal data: octave:20> r = randn(2,3,4) r = ans(:,:,1) = -1.05765 -0.16023 0.72208 0.48319 0.84360 0.20065

ans(:,:,2) = -0.31952 -0.79753 1.01204 0.53750 -0.94572 -0.64897

ans(:,:,3) = -1.62063 1.80322 0.59228 0.42084 1.21434 2.09797

ans(:,:,4) = -1.724684 0.537532 -1.094533 0.786163 -0.568425 -0.043409

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

10

Octave – Matrix Operations • Example of a Hilbert matrix and its inverse: octave:29>H hilb(4) = H= 1.00000 0.50000 0.33333 0.25000 0.50000 0.33333 0.25000 0.20000 0.33333 0.25000 0.20000 0.16667 0.25000 0.20000 0.16667 0.14286

octave:30>H_inv = inv(H) H inv =   _ 16.000 -120.000 240.000 -140.000 -120.000 1200.000 -2700.000 1680.000 240.000 -2700.000 6480.000 -4200.000 -140.000 1680.000 -4200.000 2800.000

_ octave:31>H*H inv ans = 1.0000e+00 5.6843e-14 2.2737e-13 1.1369e-13 - 1.6209e-14 1.0000e+00 -1.6032e-13 1.4477e-13 1.3137e-14 3.4139e-15 1.0000e+00 3.0947e-14 9.7387e-15 1.9165e-14 -1.1285e-13 1.0000e+00

Distribution Statement A: Cleared for Public Release. Distribution is unlimited.

11