Computer Science 1510
Lecture 9
September 26, 2016
Lecture Outline
Repetition:
DO
DO WHILE
DO EXIT
Examples involving selection and repetition
CS1510 Lecture 9
Repetition
With IF and SELECT-CASE statements we saw how
to execute portions of code only if some condition
is satisfied.
On the other hand, in some instances we may want
to execute a section of code many times.
The Fortran construct that can accomplish this
repetition or iteration of statements is called a DO
loop.
The DO loop has 3 different forms:
1. The counter-controlled DO loop which iterates
some specified number of times.
2. The DO WHILE loop which iterates while some
condition is true.
3. The DO EXIT loop which is a generalized version
of a DO WHILE loop.
CS1510 Lecture 9 1
Counter-controlled DO loop
Syntax:
DO control-variable=initial,limit,step-size
statements-to-repeat
END DO
Statements in between DO and END DO, referred to
as the body of the loop, are repeatedly executed.
The integer control-variable is referred to as
the loop counter.
Execution proceeds as follows:
1. Set the value of control-variable to initial.
2. Check to see if control-variable is
limit if step-size > 0, or
limit if step-size < 0.
3. If so, the body of the loop is executed, step-size
is added to control-variable and step 2 is
repeated. Otherwise, repetition terminates.
If not specified, the value of step-size is one.
CS1510 Lecture 9 2
Example 1: DO
PROGRAM Hello
INTEGER::i
DO i=1,10
WRITE(*,*) Hello
END DO
END PROGRAM Hello
The above code prints Hello 10 times.
Execution proceeds as follows:
1. The loop counter i is initialized to 1.
2. Since this value is less than the limit of 10, the
WRITE statement is executed.
3. We go back to the DO statement where i is
increased by one (the default value).
4. This value is less than 10 so the WRITE
statement is executed.
5. This iteration continues up to i=10, when the
WRITE statement is executed for the last time
since i is then increased to 11 which is greater
than 10, ending the loop.
6. Execution moves to the statement following the
END DO.
CS1510 Lecture 9 3
Example 2: DO
PROGRAM Odd
INTEGER::i,n
READ(*,*) n
DO i=1,n,2
WRITE(*,*) i
END DO
END PROGRAM Odd
Prints out the odd numbers less than or equal to n.
The loop counter i is initialized to 1, and is
increased by 2 on each iteration of the loop. The
loop terminates when i is greater than n.
If n is odd, say 7, then i will have the value 1 on
the first iteration of the loop, 3 on the second, 5
on the third, and 7 on the fourth, after which the
loop terminates. Thus, the WRITE statement was
executed 4 times.
If n is even, say 8, then the loop is again executed 4
times, with the same output as above since 7+2=9
(which is >8).
CS1510 Lecture 9 4
Example 3: DO
What is the value of the loop counter at the end of
the loop execution?
For example, if we were to print the value of i after
the loop in the first example:
PROGRAM Hello
INTEGER::i
DO i=1,10
WRITE(*,*) Hello
END DO
WRITE(*,*) i = ,i
END PROGRAM Hello
we would obtain i=11 since the counter is increased
prior to checking if it has exceeded the limit.
CS1510 Lecture 9 5
Example 4: DO
The only restriction on the value of step-size
is that it be nonzero, therefore, we can have
a negative step-size such that the value of
control-variable is decremented.
Repetition continues as long as the value of
control-variable is greater than or equal to
limit.
Example:
PROGRAM Hello
INTEGER::i
DO i=10,1,-1
WRITE(*,*) Hello
END DO
WRITE(*,*) i = ,i
END PROGRAM Hello
Hello would be printed 10 times and the value of
i printed after the loop would be 0.
CS1510 Lecture 9 6
Counter-controlled DO loop
The number of repetitions of a counter-controlled
DO loop is determined prior to the start of repetition.
This number depends on the values of init, limit,
and step-size.
Although the values of variables init, limit, and
step-size can be changed within the body of the
loop, such a change does not affect the number of
repetitions.
Attempting to change the value of the
control-variable within the body of the loop
will result in a compile-time error.
CS1510 Lecture 9 7
Example 5: DO
The initial value, the limit, and the step-size can be
variables or expressions in addition to constants.
Example:
PROGRAM Sum_of_integers
IMPLICIT NONE
INTEGER :: num, i, sum=0
WRITE(*,*) This program prints the sum &
& 1 + 2 + 3 +...+ num
WRITE(*,*) Enter a value for num
READ(*,*) num
DO i=1,num
sum = sum + i
END DO
WRITE(*,*) 1 + 2 + 3 +...+ ,num, = ,sum
END PROGRAM Sum_of_integers
CS1510 Lecture 9 8
Example 6: DO
DO loops can be nested.
Example:
PROGRAM Mult_table
IMPLICIT NONE
INTEGER :: m, n, lastm, lastn, prod
WRITE(*,*) Calculating m*n up to some limit
WRITE(*,*) Enter the limit of m and n
READ(*,*) lastm, lastn
WRITE(*,*) M N M*N
WRITE(*,*) -------------
DO m=1,lastm
DO n=1,lastn
prod = m*n
WRITE(*,2) m,n,prod
END DO
END DO
2 FORMAT(I2,2X,I2,2X,I3)
END PROGRAM Mult_table
CS1510 Lecture 9 9
Fortran statements: DO-WHILE
A DO-WHILE loop combines the iteration of a
DO loop with the conditional execution of an IF
statement.
This is useful when the number of repetitions is not
known in advance.
Syntax:
DO WHILE (logical-expression)
statements-to-repeat
END DO
logical-expression is any expression that
evaluates to true or false.
The loop iterates as long as logical-expression
is true.
CS1510 Lecture 9 10
Example 1: DO WHILE
PROGRAM While
IMPLICIT NONE
INTEGER::n,m
n=0
m=5
DO WHILE((m-n)>0)
n=n+1
WRITE(*,90) m,n
END DO
90 FORMAT(m = ,I2, n = ,I2)
END PROGRAM While
Output:
m = 5 n = 1
m = 5 n = 2
m = 5 n = 3
m = 5 n = 4
m = 5 n = 5
CS1510 Lecture 9 11
Example 2: DO WHILE
PROGRAM While_odd
INTEGER::i,n
i=1
READ(*,*) n
DO WHILE (i<=n)
WRITE(*,*) i
i=i+2
END DO
END PROGRAM While_odd
Like PROGRAM Odd, the above program prints the
odd numbers less than or equal to n, but uses a
DO WHILE loop instead of a counter-controlled DO
loop.
CS1510 Lecture 9 12
Fortran statements: DO EXIT
Syntax:
DO
statement-sequence-1
IF (logical-expression) EXIT
statement-sequence-2
END DO
IF logical-expression is true, the EXIT
command causes the execution to break out of
the loop, that is, execution is immediately moved
to the statement following the END DO.
A DO EXIT loop behaves similar to a DO WHILE
loop in the case where statement-sequence-1 is
not present.
One has to be careful to not introduce an infinite
loop (ie. a loop that never stops iterating)
since termination of a DO EXIT loop requires
logical-expression to be true at some point.
CS1510 Lecture 9 13
Example: DO EXIT
PROGRAM Summation
IMPLICIT NONE
INTEGER :: num, sum, limit
WRITE(*,*) Finding smallest value of n such that &
& 1+2+...+n exceeds limit
WRITE(*,*) Enter value for limit
READ(*,*) limit
num = 0
sum = 0
DO
IF (sum > limit) EXIT
num = num + 1
sum = sum + num
END DO
WRITE(*,*) 1+...+,num,=,sum,>,limit
END PROGRAM Summation
CS1510 Lecture 9 14
Fortran statements: CYCLE
In some cases we may want to terminate the current
iteration of a loop and return to the beginning of
the loop for the next iteration.
This can be accomplished by using a CYCLE
statement.
Syntax:
DO i=1,N
statement-sequence-1
IF (logical-expression-1) CYCLE
statement-sequence-2
END DO
If logical-expression-1 is true on any given
iteration, statement-sequence-2 is skipped, and
execution returns to the top of the loop.
A CYCLE statement can be used in any type of DO
loop.
CS1510 Lecture 9 15
A few more points about loops
The statements within each loop should be indented
for clarity.
It is possible for the body of a loop to never be
executed. For example, in a counter-controlled DO
loop, if initial exceeds limit on the first check,
then control jumps to the statement following the
END DO.
EXIT and CYCLE can also be used in counter-
controlled DO and DO WHILE loops.
CS1510 Lecture 9 16
Program 1: Mean time to failure
PROGRAM Mean_Time_to_Failure
!-----------------------------------------------------------------------
! Program to read a list of failure times, count them, and find the
! mean time to failure. Values are read until an end-of-data flag
! is read. Identifiers used are:
! INPUT:
! A list of failure times (FailureTime = current failure time read)
! OUTPUT:
! NumTimes : the number of failure time readings
! MeanFailureTime : the mean time to failure
!-----------------------------------------------------------------------
IMPLICIT NONE
INTEGER :: NumTimes
REAL :: FailureTime, Sum, MeanFailureTime
REAL, PARAMETER :: EndDataFlag = -1.0
Sum=0.0
NumTimes=0
WRITE(*,*) "Enter failure time of", EndDataFlag, "to stop."
DO
WRITE(*,*) "Enter failure time:"
READ(*,*) FailureTime
! If end-of-data, terminate repetition
IF (FailureTime == EndDataFlag) EXIT
NumTimes = NumTimes + 1
Sum = Sum + FailureTime
END DO
IF (NumTimes /= 0) THEN
MeanFailureTime = Sum / NumTimes
WRITE(*,*)
WRITE(*,*) "Number of failure time readings:", NumTimes
WRITE(*,*) "Mean time to failure:", MeanFailureTime
ELSE
WRITE(*,*) "No failure times were entered."
END IF
END PROGRAM Mean_Time_to_Failure
CS1510 Lecture 9 17
Example 2: Approximating sin(x) and cos(x)
Both the sine and cosine functions are represented
by infinite series:
X
x2n+1 x3 x5 x7
n
sin(x) = (1) = x + +
n=0
(2n + 1)! 3! 5! 7!
X
x 2n
x 2
x 4
x 6
cos(x) = (1)n = 1 + +
n=0
(2n)! 2! 4! 6!
Thus, we can approximate sin(x) or cos(x) for some
x by computing some finite number of terms in the
corresponding sum, call this number of terms n.
The following program approximates the value of
sin(x) using the above sum.
CS1510 Lecture 9 18
Program 2: Approximating sine
PROGRAM Approx_sine
IMPLICIT NONE
INTEGER::i,n
REAL::x,term,approx
WRITE(*,*) Enter a value for x
READ(*,*) x
WRITE(*,*) How many terms would you like to include?
READ(*,*) n
term=x
approx=x
DO i=1,n-1
term=(-1)*term*(x**2/(2*i*(2*i+1)))
approx=approx+term
END DO
WRITE(*,*) sin(,x,) is approximately,approx
END PROGRAM Approx_sine
CS1510 Lecture 9 19
Program 3: Approximating sine/cosine
PROGRAM Functions
IMPLICIT NONE
INTEGER::i,n
REAL::x,term,approx
CHARACTER::func
WRITE(*,*) Enter s for sin, and c for cos or q to quit
READ(*,*) func
DO WHILE(func/=q)
WRITE(*,*) Enter a value for x
READ(*,*) x
WRITE(*,*) How many terms would you like to include?
READ(*,*) n
SELECT CASE (func)
CASE(s)
term=x
approx=x
DO i=1,n-1
term=(-1)*term*(x**2/(2*i*(2*i+1)))
approx=approx+term
END DO
WRITE(*,*) sin(,x,) is approximately,approx
CASE(c)
term=1
approx=1
DO i=1,n-1
term=(-1)*term*(x**2/(2*i*(2*i-1)))
approx=approx+term
END DO
WRITE(*,*) cos(,x,) is approximately,approx
CASE DEFAULT
WRITE(*,*) Invalid entry, try again
END SELECT
WRITE(*,*) Enter s for sin, and c for cos or q to quit
READ(*,*) func
END DO
END PROGRAM Functions
CS1510 Lecture 9 20
Example 3: Euclids algorithm
Suppose that you have two integers m and n, and
you want to find their greatest common divisor
(GCD).
According to Euclids algorithm we do the following.
Taking m = 1976 and n = 1032,
1976 = 1032 1 + 944
1032 = 944 1 + 88
944 = 88 10 + 64
88 = 64 1 + 24
64 = 24 2 + 16
24 = 16 1 + 8
16 = 8 2 + 0
Then the GCD is 8 (the final divisor).
CS1510 Lecture 9 21
Euclids algorithm - Why it works
Suppose that we want to find the GCD of a and b.
We know that a = sd and b = td for some integers
s and t, where d is the GCD.
If we divide a by b we obtain a quotient q and a
remainder r such that a = qb + r.
= r = a qb = sd qtd = (s qt)d
Thus d is also a divisor of r.
Therefore, the GCD of a and b is also the GCD of
b and r.
Since r < b we will eventually obtain r = 0.
CS1510 Lecture 9 22
Program 4: Euclids algorithm
PROGRAM Euclid
IMPLICIT NONE
INTEGER::m,n,q,r
WRITE(*,*) Enter two integers m and n, where m>n
READ(*,*) m,n
r=1
DO WHILE (r/=0)
q=m/n
r=m-q*n
WRITE(*,*) m, = ,n, * ,q, + ,r
m=n
n=r
END DO
WRITE(*,*) The greatest common divisor is ,m
END PROGRAM Euclid
CS1510 Lecture 9 23
Example 4: Linear Least-Squares
Suppose that you have data collected from an
experiment, and you want to compute a line of
best fit.
We want to compute values for the slope m, and
the y-intercept b of a line y = mx + b that in some
way best fits our data (m and b are called regression
coefficients).
In the method of least squares, the best fit is
obtained by minimizing the sum of the squares of
the deviations of the observed y-values from the
predicted y-values,
X
n
[yi (mxi + b)]2
i=1
.
Minimizing this function gives,
P P
( xy) ( x) y
m= P 2 P b = y m
x
( x ) ( x) x
where x
and y are the mean values.
CS1510 Lecture 9 24
Program 5: Linear Least-Squares
PROGRAM Least_Squares_Line
IMPLICIT NONE
INTEGER :: NumPoints
REAL :: X, Y, Sum_of_X, Sum_of_X_squared, Sum_of_Y, Sum_of_XY,&
X_Mean, Y_Mean, Slope, Y_Intercept
REAL, PARAMETER :: EndDataFlag = -99.0
NumPoints = 0
Sum_of_X = 0.0
Sum_of_X_squared = 0.0
Sum_of_Y = 0.0
Sum_of_XY = 0.0
WRITE(*,*) "To stop, enter", EndDataFlag, " for coordinates of point."
DO
WRITE (*, (1X, A), ADVANCE = "NO") "Enter point: "
READ(*,*) X, Y
! If end of data, terminate repetition.
IF ((X == EndDataFlag) .OR. (Y == EndDataFlag)) EXIT
NumPoints = NumPoints + 1
Sum_of_X = Sum_of_X + X
Sum_of_X_squared = Sum_of_X_squared + X ** 2
Sum_of_Y = Sum_of_Y + Y
Sum_of_XY = Sum_of_XY + X * Y
END DO
! Find equation of least-squares line
X_Mean = Sum_of_X / NumPoints
Y_Mean = Sum_of_Y / NumPoints
Slope = (Sum_of_XY - Sum_of_X * Y_Mean) / &
(Sum_of_X_squared - Sum_of_X * X_Mean)
Y_Intercept = Y_Mean - Slope * X_Mean
WRITE(*,*) "Equation of least-squares line is y =", &
Slope, "x +", Y_Intercept
END PROGRAM Least_Squares_Line
CS1510 Lecture 9 25
