Fortran: Genel Bilgiler

Dr. Mürtezaoğlu

ENF 102 Bilgisayar Programlama Dili

Fortran 90

Introduction

• Introduction•

• What is a Computer? • What is Hardware and Software? • Telling a Computer What To Do • Some Basic Terminology • How Does Computer Memory Work? • Numeric Storage • Programming Languages • High-level Programming Languages • An Example Problem • An Example Program • Analysis of Program • A Closer Look at the Specification Part • A Closer Look at the Execution Part • How to Write a Computer Program • A Quadratic Equation Solver - The Algorithm • A Quadratic Equation Solver - The Program • A Quadratic Equation Solver - The Testing Phase • Points Raised • Bugs - Compile-time Errors • Bugs - Run-time Errors • Compiler Switches

What is a Computer?A simple computer may look like this:

                                                                     

•memory (RAM) -- used to store values during execution of a program, •CPU (Central Processor Unit) -- does the `work', •disc drive -- `permanently' stores files, •keyboard -- allows user to input information, •VDU -- visually outputs data,

Introduction

Introduction

– What is Hardware and Software?– A computer system comprises hardware and software. – Hardware is the physical medium, for example: – circuit boards – processors – keyboard – Software are computer programs, for example: – operating system – editor – compilers – a Fortran 90 program

Introduction

• Telling a Computer What To Do• To get a computer to perform a specific task it must be given a

sequence of unambiguous instructions or a program. • We meet many examples of programs in everyday life, for

example, instructions on how to assemble a bedside cabinet. These instructions are generally numbered, meaning that there is a specific order to be followed, they are also (supposed to be) precise so that there is no confusion about what is intended:

• insert the spigot into hole `A', • apply glue along the edge of side panel, • press together side and top panels • attach toggle pin `B' to gromit `C' • ... and so on • If these instructions are not followed `to the letter', then the

cabinet would turn out wonky.

Introduction

• Some Basic Terminology• It is necessary to cover some terminology. Hopefully, much of it will be

familiar -- you will hear many of the terms used throughout the course. • Bit is short for Binary Digit. Bits have value of 1 or 0, (or on or off, or, true or

false), • 8 Bits make up 1 Byte, 1024 Bytes make up 1 KByte (1 KiloByte or 1K),

(``Why 1024?'' I hear you ask. Because . • 1024 KBytes make up 1 MByte (1 MagaByte or 1M), • 1024 MBytes make up 1 GByte (1 GigaByte or 1G), • all machines have a wordsize -- a fundamental unit of storage, for example,

8-bits, 16-bits, etc. The size of a word (in Bytes) differs between machines. A Pentium based machine is 32-bit.

• a flop is a floating point operation per second. A floating point operation occurs when two real numbers are added. Today, we talk of megaflops or even gigaflops.

• parallel processing occurs when two or more CPUs work on solution of the same problem at the same time.

Introduction

• How Does Computer Memory Work?• Here the wordsize is 8-bits:

• A computers memory is addressable, • each memory location will contain some sort of `value', • Each location has a specific `number' (represented as hexadecimal

[base-16], e.g., 3F2C), • Fortran 90 allows (English) names to be given to memory locations, • the value of a location can be read from or written to. • The CPU can say, `fetch the contents of memory location 3F2C' or

`write this value to location 3AF7'.

Introduction

• Numeric Storage• In general, there are two types of numbers used in Fortran 90

programs INTEGER s (whole numbers) and REAL s (floating point numbers).

• INTEGER s are stored exactly, often in range (-32767, 32767). • REAL s are stored approximately.

– They are partitioned into a mantissa and an exponent, – Exponent can only take a small range of values.

• You can get numeric exceptions: • overflow -- exponent is too big, • underflow -- exponent is too small. • In Fortran 90 you can decide what numeric range is to be

supported. • CHARACTER s are stored differently.

Introduction

• Programming Languages• Programming languages must be: • totally unambiguous (unlike natural languages, for

example, English), • expressive -- it must be fairly easy to program common

tasks, • practical -- it must be an easy language for the compiler

to translate, • simple to use. • All programming languages have a very precise syntax

(or grammar). • This ensures all syntactically-correct programs have a

single meaning.

Introduction

• High-level Programming Languages• Assembler code is a Low-Level Language. • Fortran 90, FORTRAN 77, ADA, C and Java are

High-Level Languages. • a program is a series of instructions to the CPU, • could write all programs in assembler code but

this is a slow, complex and error-prone process, • high-level languages are more expressive, more

secure and quicker to use, • the high-level program is compiled (translated)

into assembler code by a compiler.

Introduction

• To convert from F (Fahrenheit) to C (Centigrade) we can use the following formula:

• • To convert from C to K (Kelvin)

we add 273. • The program would accept a

Fahrenheit temperature as input and produce the Centigrade and Kelvin equivalent as output.

PROGRAM Temp_ConversionIMPLICIT NONE INTEGER :: Deg_F,

Deg_C, KPRINT*, "Please type in the temp in F“READ*, Deg_F Deg_C = 5*(Deg_F-

32)/9PRINT*, "This is equal to", Deg_C, "C"

K = Deg_C + 273 PRINT*, "and", K, "K"

END PROGRAM Temp_Conversion This program, called Temp.f90, can be

compiled: chad2-13{adamm} 26> f90 Temp.f90

NAg Fortran 90 compiler v2.2. New Debugger: 'dbx90' and run:

chad2-13{adamm} 27> a.out Please type in the temp in F 45 This is equal to 7 C and 280 K

Introduction

• Analysis of Program• The code is delimited by PROGRAM ... END

PROGRAM statements. Between these there are two distinct areas.

• Specification Part – specifies named memory locations (variables) for

use, – specifies the type of the variable,

• Execution Part – reads in data, – calculates the temp in C and K and – prints out results.

Introduction

• A Closer Look at the Specification Part• IMPLICIT NONE -- this should always be present. Means all

variables must be declared. • INTEGER :: Deg_F, Deg_C, K -- declares three INTEGER (whole

number) variables. Other variable types: – REAL -- real numbers, e.g., 3.1459, , – LOGICAL -- take vaules .TRUE. or .FALSE., – CHARACTER -- contains single alphanumeric character, e.g., 'a', – CHARACTER(LEN=12) -- contains 12 alphanumeric characters, a

string, • Fortran 90 is not case sensitive. • K is the same as k and INTEGER is the same as integer.

Introduction

• A Closer Look at the Execution Part• This is the part of the program that does the actual `work'. • PRINT*, "Please type in the temp in F" -- writes the string to the

screen, • READ*, Deg_F -- reads a value from the keyboard and assigns it to

the INTEGER variable Deg_F, • Deg_C = 5*(Deg_F-32)/9 -- the expression on the RHS is

evaluated and assigned to the INTEGER variable Deg_C, – * is the multiplication operator, – - is the subtraction operator, – / is the division operator, (takes longer than *) – = is the assignment operator.

• PRINT*, "This is equal to", Deg_C, "C" -- displays a string on the screen followed by the value of a variable (Deg_C) followed by a second string ("C").

• By default, input is from the keyboard and output to the screen.

Introduction

• How to Write a Computer Program• There are 4 main steps: • specify the problem, • analyse and break down into a series of steps

towards solution, • write the Fortran 90 code, • compile and run (i.e., test the program). • It may be necessary to iterate between steps 3

and 4 in order to remove any mistakes. • The testing phase is very important.

Introduction

•   • A Quadratic Equation Solver - The Algorithm• The problem Write a program to calculate the roots of a quadratic

equation of the form: • • The roots are given by the following formula • • The algorithm • READ values of a, b and c, • if a is zero then stop as we do not have a quadratic, • calculate value of discriminant • if D is zero then there is one root: , • if D is > 0 then there are two real roots: and , • if D is < 0 there are two complex roots: and , • PRINT solution.

Introduction• A Quadratic Equation Solver - The Program• PROGRAM QES• IMPLICIT NONE INTEGER :: a, b, c, D • REAL :: Real_Part, Imag_Part • PRINT*, "Type in values for a, b and c" • READ*, a, b, c • IF (a /= 0) THEN ! Calculate discriminant • D = b*b - 4*a*c IF (D == 0) THEN ! one root • PRINT*, "Root is ", -b/(2.0*a) ELSE IF (D > 0) THEN ! real roots • PRINT*, "Roots are",(-b+SQRT(REAL(D)))/(2.0*a),& "and",

(-b-SQRT(REAL(D)))/(2.0*a) ELSE ! complex roots • Real_Part = -b/(2.0*a)• ! D < 0 so must take SQRT of -D • Imag_Part = (SQRT(REAL(-D))/(2.0*a)) • PRINT*, "1st Root", Real_Part, "+", Imag_Part, "i" • PRINT*, "2nd Root", Real_Part, "-", Imag_Part, "i" • END IF • ELSE ! a == 0 PRINT*, "Not a quadratic equation" • END IF • END PROGRAM QES

Introduction• A Quadratic Equation Solver - The Testing Phase• The output from the program is as follows: • uxa{adamm} 35> a.out • Type in values for a, b and c • 1 -3 2 Roots are 2.0000000 and 1.0000000 • uxa{adamm} 36> a.out • Type in values for a, b and c • 1 -2 1 Root is 1.0000000 • uxa{adamm} 37> a.out • Type in values for a, b and c • 1 1 1 • 1st Root -0.5000000 + 0.8660254 i • 2nd Root -0.5000000 - 0.8660254 i • uxa{adamm} 38> a.out • Type in values for a, b and c • 0 2 3 Not a quadratic equation

• Its can be seen that the `test data' used above exercises every line of the program. This is important in demonstrating correctness.

Introduction

• Points Raised• The previous program introduces some new ideas, • comments -- anything on a line following a ! is ignored, • & -- means the line is continues, • IF construct -- different lines are executed depending on the value

of the Boolean expression, • relational operators -- == (`is equal to') or > (`is greater than'), • nested constructs -- one control construct can be located inside

another. • procedure call -- SQRT(X) returns square root of X. • type conversion -- in above call, X must be REAL. In the program,

D is INTEGER, REAL(D) converts D to be real valued. • To save CPU time we only calculate the discriminant, D, once.

Introduction

• Bugs - Compile-time Errors• In previous program, what if we accidentally typed: • Rael_Part = -b/(2.0*a) • the compiler generates a compile-time or syntax error: • uxa{adamm} 40> f90 Quad.f90 • NAg Fortran 90 compiler v2.2. • Error: Quad.f90, line 16: • Implicit type for RAEL_PART • detected at RAEL_PART@= • Error: Quad.f90, line 24: • Symbol REAL_PART referenced but never set • detected at QES@<end-of-statement> • [f90 terminated - errors found by pass 1]

Introduction

• Bugs - Run-time Errors

• If we had typed • Real_Part = -b/(.0*a) • then the program would compile but we would get a run-

time error, • uxa{adamm} 43> a.out • Type in values for a, b and c • 1 1 1• *** Arithmetic exception: Floating divide by zero -

aborting • Abort • It is also possible to write a program that gives the wrong

results!

Introduction

• Compiler Switches• Compiler is invoked by f90 • f90 Quad.f90 This: • checks the program syntax • generates executable code in a file a.out • Switches or flags can be supplied to compiler: • -o < output-filename >: gives executable a different

name, for example, Quad • f90 Quad.f90 -o Quad-time: report execution times

f90 Quad.f90 -o Quad –time For more information about the compiler type:

• man f90