Introduction to Scilab

Kannan M. Moudgalya
IIT Bombay
www.moudgalya.org
[email protected]

Scilab Workshop
Bhaskaracharya Pratishtana
4 July 2009

Kannan Moudgalya Introduction to Scilab 1/52

Outline

I Software engineering
I Implication to scientific computations
I Scilab as a possible solution
I Scilab - a tutorial introduction
I Simple arithmetic
I Matrix operations
I Vector arithmetic
I Conditionals
I Plots
I Installation

Kannan Moudgalya Introduction to Scilab 2/52

Software - Bottleneck

I In the past 40 years hardware technology advanced more than

1,000 times
I Software technology about same - people take same time to
write programs as before - perhaps twice or thrice faster, with
good editors, etc.
I Productivity: 5 lines/man day at IBM with 1 error/mloc
(million lines of code)
I The above includes testing, debugging and documentation
I Such a low productivity with sophisticated software tools and
that too by experienced programmers
I Others will produce even less

Kannan Moudgalya Introduction to Scilab 3/52

Shortcomings in Science Education

I Scientists and engineers often have no formal course in

programming
I Perhaps one course mainly Fortran, perhaps some C
I Objective: to solve engineering problems. Good programming
is not a must
I Computer science departments do not spend much time on
numerical programming (e.g. does not teach Fortran)
I Shortage of computer scientists
I Scientists and engineers teach programming to their students
I Many of them would have had no formal course in
programming

Kannan Moudgalya Introduction to Scilab 4/52

Gap between Computer Scientists and Engineers

I Numerical software is not of interest to computer scientists

I Computers scientists do not carry out numerical simulations
I Their programs work mostly with characters - may be integers
I This, combined with shortcomings in science education, as
mentioned in the last slide, makes scientific computations a
difficult problem

Kannan Moudgalya Introduction to Scilab 5/52

Idea to Implementation

I R&D Stage: Evaluation of ideas

I Need rapid idea testing workbench
I Come up with idea
loop
Test it
Modify it
end loop
I When idea works, go to production phase:
I Convert code for repeated use
I Use code repeatedly
I Need efficient code

Kannan Moudgalya Introduction to Scilab 6/52

College Education = Idea Testing

I College teaches new things

I Extensive idea testing phase
I Need a tool for idea testing
I Performance of code often does not matter

I Scilab fits the bill

Kannan Moudgalya Introduction to Scilab 7/52

Applied Research Institutions

I High productivity platform for idea testing

I Developing efficient code for tested ideas

I Scilab is one such tool

Kannan Moudgalya Introduction to Scilab 8/52

Scilab

I Environment for numerical computer applications

I Good mathematical library in compiled C code
I Interpreted high level language
I High productivity tool Scilab:C = C:Assembly
I Can work with Fortran, C: Transition to production phase
possible
I Good graphics capability
I Large installed base
I A lot of algorithms implemented in interpreted language as
well
I Free
I Check out www.scilab.org or www.scilab.in

Kannan Moudgalya Introduction to Scilab 9/52

History of Scilab

I Idea of Cleve Moler, CS Prof. New Mexico State University

I Was in Linpack, Eispack (robust algorithms) projects
I NSF sponsored project to develop Matlab in Fortran - Free
I Many companies started using this idea
I Matrixx
I CTRL-C
I Matlab
I Scilab
I Used extensively for linear algebra, simulation, control system
design

Kannan Moudgalya Introduction to Scilab 10/52

Mathematical Library

I Special functions
I Bessel
I Gamma
I Error function
I Elliptic integral
I Polynomials
I Characteristic polynomial
I Roots
I Multiplication
I Division
I Curve fitting
I Matrix condition
I Condition number
I 1,2,F and ∞ norms
I rank

Kannan Moudgalya Introduction to Scilab 11/52

Mathematical Library - Continued
I Matrix functions
I Exponential
I Powers
I Log
I Square root
I Decomposition & factorisation
I LU
I QR
I SVD
I Cholesky
I Schur
I Inverse
I Signal processing
I FFT, FFT2, IFFT, IFFT2
I Convolution
I Deconvolution
I Correlation coefficient
Kannan Moudgalya Introduction to Scilab 12/52
Scilab’s Language

I C like langugae
I Control flow
I if
I while
I select
I break
I Procedures
I Scripts
I Functions
I Other features
I Diary
I Can call C and Fortran programs

Kannan Moudgalya Introduction to Scilab 13/52

Features of Scilab

I Scilab is made up of three distinct parts:

I An interpreter
I Libraries of functions (Scilab procedures)
I Libraries of Fortran and C routines
I It includes hundreds of mathematical functions with the
possibility to interactively add programs from various
languages (C, Fortran).
I It has sophisticated data structures including lists,
polynomials, rational functions, linear systems, etc.

Kannan Moudgalya Introduction to Scilab 14/52

How to Download Scilab?

I Scilab can be downloaded from

ftp://ftp.iitb.ac.in/misc packages/Scilab/
I The website of Scilab is www.scilab.org
I It is distributed in source code format.
I Binaries for Windows95/NT, Unix/Linux/Mac OS/X are also
available. All the binary versions include tk/tcl interface.

Kannan Moudgalya Introduction to Scilab 15/52

Usage of Scilab

Kannan Moudgalya Introduction to Scilab 16/52

Simple Arithmetic - 1

4+6+12

ans =
22.
a = 4, b = 6; c = 12

a =
4.
c =
12.
a+b+c

ans =
22.
Kannan Moudgalya Introduction to Scilab 17/52
Useful Commands

I demos
I Gives demos on several different things
I apropos
I Helps locate commands associated with a word
I help
I functional invocation with no arguments
I Helps draw plots
I diary
I Stores all commands and resulting outputs

Kannan Moudgalya Introduction to Scilab 18/52

Simple Arithmetic & Display

a = 4; b = 6; c = 12;
d = a+b+c

d =
22.

d = a+b+c;

d =

22.

Kannan Moudgalya Introduction to Scilab 19/52

Simple Arithmetic

format(’v’,10)
e = 1/30

e =
0.0333333
format(’v’,20)
e

e =
0.03333333333333333
format(’e’,20)
e

e =
3.3333333333333E-02
Kannan Moudgalya Introduction to Scilab 20/52
Simple Arithmetic

format(’v’,10)
x = sqrt(2)/2, y = asin(x)

x =

0.7071068
y =

0.7853982

y_deg = y * 180 /%pi

y_deg =

45.
Kannan Moudgalya Introduction to Scilab 21/52
Rounding, Truncation, etc.

x = 2.6, y1 = fix(x), y2 = floor(x), y3 = ceil(x), ...

y4 = round(x)

x =

2.6
y1 =

2.
y2 =

2.
y3 =

3.
y4 =
Kannan Moudgalya Introduction to Scilab 22/52
Different Ways to Specify a List

The following three commands produce identical result:

x = [0 .1*%pi .2*%pi .3*%pi .4*%pi .5*%pi .6*%pi ...

.7*%pi .8*%pi .9*%pi %pi];

x = (0:0.1:1)*%pi;

x = linspace(0,%pi,11);

Kannan Moudgalya Introduction to Scilab 23/52

Vector Operation - 1

-->x = (0:0.1:1)*%pi;

-->y = sin(x)

y =
column 1 to 6

! 0. 0.3090170 .5877853 0.8090170 0.9510565 1. !

column 7 to 11

! 0.9510565 0.8090170 0.5877853 0.3090170 1.225E-16 !

-->y(5)

ans =
0.9510565
Kannan Moudgalya Introduction to Scilab 24/52
Vector Operation - 2

-->a = 1:5, b = 1:2:9

a =
! 1. 2. 3. 4. 5. !
b =
! 1. 3. 5. 7. 9. !
-->c = [b a]

c =
! 1. 3. 5. 7. 9. 1. 2. 3. 4. 5. !
-->d = [b(1:2:5) 1 0 1]

d =
! 1. 5. 9. 1. 0. 1. !
Kannan Moudgalya Introduction to Scilab 25/52
Vector Operation - 3

-->a, b

a =
! 1. 2. 3. 4. 5. !
b =
! 1. 3. 5. 7. 9. !
-->a - 2

ans =
! - 1. 0. 1. 2. 3. !
-->2*a-b

ans =
! 1. 1. 1. 1. 1. !
Kannan Moudgalya Introduction to Scilab 26/52
Vector Operation - 5

-->a

a =
! 1. 2. 3. 4. 5. !

-->a.^2

ans =
! 1. 4. 9. 16. 25. !

-->a.^a

ans =
! 1. 4. 27. 256. 3125. !
Kannan Moudgalya Introduction to Scilab 27/52
Vector Operation - 6

-->a, b

a =
! 1. 2. 3. 4. 5. !
b =
! 1. 3. 5. 7. 9. !

-->a./b

ans =
! 1. 0.6666667 0.6 0.5714286 0.5555556 !

-->b.\a

ans =
! 1. 0.6666667 0.6 0.5714286 0.5555556 !
Kannan Moudgalya Introduction to Scilab 28/52
Vector Operation - 7

-->a, b

a =
! 1. 2. 3. 4. 5. !
b =
! 1. 3. 5. 7. 9. !

-->a/b

ans =
0.5757576

Kannan Moudgalya Introduction to Scilab 29/52

Vector Operation - 8

-->a, b

a =
! 1. 2. 3. 4. 5. !
b =
! 1. 3. 5. 7. 9. !

-->a\b

ans =
! 0. 0. 0. 0. 0. !
! 0. 0. 0. 0. 0. !
! 0. 0. 0. 0. 0. !
! 0. 0. 0. 0. 0. !
! 0.2 0.6 1. 1.4 1.8 !
Kannan Moudgalya Introduction to Scilab 30/52
Machine Epsilon

-->num=0; EPS=1;

-->while (1+EPS)>1
--> EPS = EPS/2;
--> num = num+1;
-->end

-->num

num =
53.

-->EPS=2*EPS

EPS =
2.220E-16
Kannan Moudgalya Introduction to Scilab 31/52
Logical Operators

== equal to
< less than
> greater than
<= less than or equal to
>= greater than or equal to
<> or ∼= not equal to

Kannan Moudgalya Introduction to Scilab 32/52

Use of Machine Epsilon

-->x = (-2:2)/3

x =
! - 0.6666667 - 0.3333333 0. 0.3333333 0.6666667 !

-->sin(x)./x

!--error 27
division by zero...

Kannan Moudgalya Introduction to Scilab 33/52

Use of Machine Epsilon

-->x = x+(x==0)*%eps

x =
! -0.6666667 -0.3333333 2.22E-16 0.3333333 0.6666667 !

-->sin(x)./x

ans =
! 0.9275547 0.9815841 1. 0.9815841 0.9275547 !

Kannan Moudgalya Introduction to Scilab 34/52

Vector Operations Using Logical Operators

-->A = 1:9, B = 9-A

A =
! 1. 2. 3. 4. 5. 6. 7. 8. 9. !
B =
! 8. 7. 6. 5. 4. 3. 2. 1. 0. !
-->tf = A==B

tf =
! F F F F F F F F F !
-->tf = A>B

tf =
! F F F F T T T T T !
Kannan Moudgalya Introduction to Scilab 35/52
Transpose

-->c = [1;2;3]

c =
! 1. !
! 2. !
! 3. !
-->a=1:3

a =
! 1. 2. 3. !
-->b = a’

b =
! 1. !
! 2. !
! 3. !
Kannan Moudgalya Introduction to Scilab 36/52
Submatrix

-->A=[1 2 3;4 5 6;7 8 9]

A =

! 1. 2. 3. !
! 4. 5. 6. !
! 7. 8. 9. !

-->A(3,3)=0

A =

! 1. 2. 3. !
! 4. 5. 6. !
! 7. 8. 0. !
Kannan Moudgalya Introduction to Scilab 37/52
Submatrix

A =

! 1. 2. 3. !
! 4. 5. 6. !
! 7. 8. 0. !

-->B=A(3:-1:1,1:3)

B =
! 7. 8. 0. !
! 4. 5. 6. !
! 1. 2. 3. !

Kannan Moudgalya Introduction to Scilab 38/52

Submatrix

-->A

A =
! 1. 2. 3. !
! 1. 4. 7. !
! 7. 8. 0. !

-->B=A(:,2)

B =
! 2. !
! 4. !
! 8. !

Kannan Moudgalya Introduction to Scilab 39/52

Submatrix

-->b=[5 -3;2 -4]

b =
! 5. - 3. !
! 2. - 4. !
-->x=abs(b)>2

x =
! T T !
! F T !
-->y=b(abs(b)>2)

y
=
! 5. !
! - 3. !
! - 4. !
Kannan Moudgalya Introduction to Scilab 40/52
Special Matrices

-->zeros(3,3)

ans =
! 0. 0. 0. !
! 0. 0. 0. !
! 0. 0. 0. !
-->ones(2,4)

ans =
! 1. 1. 1. 1. !
! 1. 1. 1. 1. !
-->rand(2,1)

ans =
! 0.2113249 !
! 0.7560439 !
Kannan Moudgalya Introduction to Scilab 41/52
Go for Vector Computation

Kannan Moudgalya Introduction to Scilab 42/52

Go for Vector Computation

-->a = ones(10000,1);
-->timer()

ans =
0.02
-->for i = 1:10000, b(i)=a(i)+a(i); end
-->timer()

ans =
0.31
-->c = a+a;
-->timer()

ans =
0.03
Kannan Moudgalya Introduction to Scilab 43/52
Plots

Go through the Demos!

Kannan Moudgalya Introduction to Scilab 44/52

1 t = ( 0 : 0 . 1 : 6 ∗ %pi ) ;
2 plot2d ( t ’ , s i n ( t ) ’ ) ;
3 x t i t l e ( ’ p l o t 2 d and x g r i d ’ , ’ t ’ , ’ s i n ( t ) ’ ) ;
4 xgrid ( ) ;

Kannan Moudgalya Introduction to Scilab 45/52

1 plot2d1 ( ’ e n l ’ , 1 , ( 1 : 1 0 : 1 0 0 0 0 ) ’ ) ;
2 x t i t l e ( ’ plot2d1 log scale ’ , ’ t ’ , ’ y log scale ’ );
3 xgrid ( 3 ) ;

Kannan Moudgalya Introduction to Scilab 46/52

1 subplot ( 2 ,2 ,1); plot3d ( ) ;
2 subplot ( 2 ,2 ,2); plot2d ( ) ;
3 subplot ( 2 ,2 ,3); histplot ();
4 subplot ( 2 ,2 ,4); grayplot ( ) ;

Kannan Moudgalya Introduction to Scilab 47/52

1 plot3d ( ) ;
2 T i t l e =[ ’ p l o t 3 d : z=s i n ( x ) ∗ c o s ( y ) ’ ] ;
3 xtitle ( Title , ’ ’ , ’ ’ );

Kannan Moudgalya Introduction to Scilab 48/52

System requirements

I Source version
I Scilab requires approximately 130 MB of disk storage to
unpack and install (all sources included).
I Also, X Window (X11R4, X11R5 or X11R6), C and Fortran
compilers are needed.
I Binary version
I The minimum requirement for running Scilab (without
sources) is about 40 MB when decompressed.
I Being partially and statically linked, these versions do not
require a Fortran compiler.

Kannan Moudgalya Introduction to Scilab 49/52

How to install Scilab?

I Windows
I Download scilab-4.1.exe
I Click this file and follow the instructions
I Launch from its icon on the Desktop.
I Linux
I Download scilab-4.1.bin.linux-i686.tar.gz
I Issue the following commands
I tar zxvf scilab-4.1.bin.linux-i686.tar.gz
I cd scilab-4.1
I make
I The binary is at bin/scilab

Kannan Moudgalya Introduction to Scilab 50/52

Conclusions

I Scilab is ideal for educational institutions, including schools

I Built on a sound numerical platform
I It is free
I Also suitable for industrial applications
I Standard tradeoff between free and commercial applications

Kannan Moudgalya Introduction to Scilab 51/52

 Thank you

Kannan Moudgalya Introduction to Scilab 52/52