MAT 685:

C++ for Math-

ematicians

John Perry

Background

Implementa-
tion
Passing by reference
MAT 685: C++ for Mathematicians
Overloading
Implementation The Extended Euclidean Algorithm
Summary

John Perry

University of Southern Mississippi

Spring 2017
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
Bézout’s Identity
John Perry

Background

Implementa-
tion Theorem (Bézout’s Identity)
Passing by reference
Overloading Let a, b ∈ Z. We can find x, y ∈ Z such that
Implementation

Summary
gcd (a, b) = ax + by .

Moreover, gcd (a, b) is the smallest positive integer of the form

ax + by where x, y ∈ Z.
 MAT 685:
C++ for Math-
ematicians
Bézout’s Identity
John Perry

Background

Implementa-
tion Theorem (Bézout’s Identity)
Passing by reference
Overloading Let a, b ∈ Z. We can find x, y ∈ Z such that
Implementation

Summary
gcd (a, b) = ax + by .

Moreover, gcd (a, b) is the smallest positive integer of the form

ax + by where x, y ∈ Z.

Examples
• gcd (24, −16) = 8 = 24 × 1 + (−16) × 1
• gcd (24, −15) = 3 = 24 × 2 + × (−15) × 3
 MAT 685:
C++ for Math-
ematicians
Extended Euclidean Algorithm
John Perry

Background

Implementa-
Recall
tion If c = a mod b then gcd (a, b) = gcd (b, c).
Passing by reference
Overloading
Implementation

Summary
 MAT 685:
C++ for Math-
ematicians
Extended Euclidean Algorithm
John Perry

Background

Implementa-
Recall
tion If c = a mod b then gcd (a, b) = gcd (b, c).
Passing by reference
Overloading
Implementation
Modify recursive Euclidean Algorithm:
Summary given
• a, b ∈ Z arguments
• c = a mod b from recursion
• u, v ∈ Z s.t. gcd (b, c) = bu + cv from recursion
 MAT 685:
C++ for Math-
ematicians
Extended Euclidean Algorithm
John Perry

Background

Implementa-
Recall
tion If c = a mod b then gcd (a, b) = gcd (b, c).
Passing by reference
Overloading
Implementation
Modify recursive Euclidean Algorithm:
Summary given
• a, b ∈ Z arguments
• c = a mod b from recursion
• u, v ∈ Z s.t. gcd (b, c) = bu + cv from recursion
choose q: a = qb + c compute
rewrite: c = a − qb theory
 MAT 685:
C++ for Math-
ematicians
Extended Euclidean Algorithm
John Perry

Background

Implementa-
Recall
tion If c = a mod b then gcd (a, b) = gcd (b, c).
Passing by reference
Overloading
Implementation
Modify recursive Euclidean Algorithm:
Summary given
• a, b ∈ Z arguments
• c = a mod b from recursion
• u, v ∈ Z s.t. gcd (b, c) = bu + cv from recursion
choose q: a = qb + c compute
rewrite: c = a − qb theory
substitute: gcd (b, c) = bu + (a − qb)v theory
rewrite: gcd (b, c) = av + b (u − qv) theory
 MAT 685:
C++ for Math-
ematicians
Extended Euclidean Algorithm
John Perry

Background

Implementa-
Recall
tion If c = a mod b then gcd (a, b) = gcd (b, c).
Passing by reference
Overloading
Implementation
Modify recursive Euclidean Algorithm:
Summary given
• a, b ∈ Z arguments
• c = a mod b from recursion
• u, v ∈ Z s.t. gcd (b, c) = bu + cv from recursion
choose q: a = qb + c compute
rewrite: c = a − qb theory
substitute: gcd (b, c) = bu + (a − qb)v theory
rewrite: gcd (b, c) = av + b (u − qv) theory
substitute: gcd (a, b) = av + b(u − qv) theory
 MAT 685:
C++ for Math-
ematicians
Extended Euclidean Algorithm
John Perry

Background

Implementa-
Recall
tion If c = a mod b then gcd (a, b) = gcd (b, c).
Passing by reference
Overloading
Implementation
Modify recursive Euclidean Algorithm:
Summary given
• a, b ∈ Z arguments
• c = a mod b from recursion
• u, v ∈ Z s.t. gcd (b, c) = bu + cv from recursion
choose q: a = qb + c compute
rewrite: c = a − qb theory
substitute: gcd (b, c) = bu + (a − qb)v theory
rewrite: gcd (b, c) = av + b (u − qv) theory
substitute: gcd (a, b) = av + b(u − qv) theory
return gcd (a, b), x = v, y = u − qv compute
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Implementation Compute gcd (26, −14):
Summary

gcd(26,-14) a=26, b=-14

• 26 ≥ 0 so no change to a
• −14 < 0 so b reassigned to 14
• a and b not 0
• c = 26 % 14 = 12
• return gcd(14,12)
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Overloading Compute gcd (26, −14):
Implementation

Summary
gcd(26,-14)
gcd(14,12) a=14, b=12
• 14 ≥ 0 so no change to a
• 12 ≥ 0 so no change to b
• a and b not 0
• c = 14 % 12 = 2
• return gcd(12,2)
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Compute gcd (26, −14):
Implementation

Summary gcd(26,-14)
gcd(14,12)
gcd(12,2) a=12, b=2
• 12 ≥ 0 so no change to a
• 2 ≥ 0 so no change to b
• a and b not 0
• c = 12 % 2 = 0
• return gcd(2,0)
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Compute gcd (26, −14):
Overloading
Implementation
gcd(26,-14)
Summary
gcd(14,12)
gcd(12,2)
gcd(2,0) a=2, b=0
• 2 ≥ 0 so no change to a
• 0 ≥ 0 so no change to b
• but b = 0
∴ return a=2
. . . cascades back up: gcd (26, −14) = 2
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Implementation
Compute gcd (26, −14):
Summary
gcd(26,-14)
gcd(14,12)
gcd(12,2)=2
gcd(2,0)=2
2=2×1+0×1
return 2, 1, 1
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Compute gcd (26, −14):
Overloading
Implementation

Summary
gcd(26,-14)
gcd(14,12)
gcd(12,2)=2
received 2, 1, 1
2=2×1+0×1
0 = 12 − 2 × 6
2 = 2 × 1 + (12 − 2 × 6) × 1
2 = 2 × (−5) + 12 × 1
return 2, 1, −5
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Compute gcd (26, −14):
Implementation

Summary gcd(26,-14)
gcd(14,12)=2
received 2, 1, −5
2 = 2 × (−5) + 12 × 1
2 = 14 − 12 × 1
2 = (14 − 12 × 1) × (−5) + 12 × 1
2 = 14 × (−5) + 12 × 6
return 2, −5, 6
 MAT 685:
C++ for Math-
ematicians
Example
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Compute gcd (26, −14):
Implementation

Summary
gcd(26,-14)=2
received 2, −5, 6
2 = 14 × (−5) + 12 × 6
12 = 26 − 14 × 1
2 = 14 × (−5) + (26 − 14 × 1) × 6
2 = 14 × (−11) + 26 × 6
2 = −14 × 11 + 26 × 6
return 2, 6, 11
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
Return 3 items?
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Implementation
We need gcd() to return gcd (a, b), x, and y
Summary
 MAT 685:
C++ for Math-
ematicians
Return 3 items?
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Implementation
We need gcd() to return gcd (a, b), x, and y
Summary

Problem
C++ procedures return only one item

Solution
Modify arguments
 MAT 685:
C++ for Math-
ematicians
Modify arguments?
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
To “return” more than one item, modify procedure’s arguments
Implementation

Summary
 MAT 685:
C++ for Math-
ematicians
Modify arguments?
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
To “return” more than one item, modify procedure’s arguments
Implementation

Summary
Problem
C++ procedures copy arguments by default
• Can change copy of argument, but. . .
• . . . doesn’t change original!

Solution
Pass by “reference”
 MAT 685:
C++ for Math-
ematicians
Pass by reference?
John Perry

Background
Listing 1: template to pass by reference
Implementa-
tion return_type proc_name(type1 & arg1, ...) {
Passing by reference
Overloading statement1;
Implementation
statement2;
Summary
...
}

Notice the difference?

 MAT 685:
C++ for Math-
ematicians
Pass by reference?
John Perry

Background
Listing 2: template to pass by reference
Implementa-
tion return_type proc_name(type1 & arg1, ...) {
Passing by reference
Overloading statement1;
Implementation
statement2;
Summary
...
}

Notice the difference?

& pass argument by reference
• argument not copied
• procedure receives actual object
• hides some “pointer magic”
(if you don’t know what pointers are, don’t
worry about it)
 MAT 685:
C++ for Math-
ematicians
C++ Interface
John Perry

Background

Implementa-
tion
Listing 3: add to recursive gcd.hpp
Passing by reference
Overloading
/**
Implementation
Calculate the greatest common divisor
Summary
of the first two integers, as well as
the Bezout coefficients.
@param a integer for gcd
@param b integer for gcd
@param x Bezout coefficient (modified)
@param y Bezout coefficient (modified)
@return gcd(a,b)
*/
long gcd(long a, long b, long & x, long & y);
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
An objection
John Perry

Background

Implementa- long gcd(long, long);

tion
Passing by reference
long gcd(long, long, long &, long &);
Overloading
Implementation

Summary Can we do this?

Already defined gcd() in gcd.hpp; can confuse computer?
 MAT 685:
C++ for Math-
ematicians
An objection
John Perry

Background

Implementa- long gcd(long, long);

tion
Passing by reference
long gcd(long, long, long &, long &);
Overloading
Implementation

Summary Can we do this?

Already defined gcd() in gcd.hpp; can confuse computer?

• Overloading
• key feature of object-oriented languages
• Simula (1967)
• Many languages do not allow this
• identifiers must be unique
• common in older languages
• Fortran, C, Pascal, . . .
 MAT 685:
C++ for Math-
ematicians
Overloading
John Perry

Background

Implementa-
tion • Same name, different procedures
Passing by reference
Overloading • Also called polymorphism
Implementation

Summary
• C++ allows w/different signature, determined by:
• number of arguments
• argument types
 MAT 685:
C++ for Math-
ematicians
Overloading
John Perry

Background

Implementa-
tion • Same name, different procedures
Passing by reference
Overloading • Also called polymorphism
Implementation

Summary
• C++ allows w/different signature, determined by:
• number of arguments
• argument types

long gcd(long, long);

long gcd(long, long, long &, long &);

• first gcd() takes two arguments

• second gcd() takes four arguments
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
C++ Implementation
John Perry

Background

Implementa-
Listing 4: add to gcd_recursive.cpp (p. 1)
tion long gcd ( long a , long b , long & x , long & y ) {
Passing by reference
Overloading
// Make s u r e a , b n o n n e g a t i v e
Implementation bool a _ n e g = f a l s e ;
bool b_neg = f a l s e ;
Summary
if ( a < 0) {
a = −a ;
a_neg = true ;
}
if ( b < 0) {
b = −b ;
b_neg = t r u e ;
}

// i f a , b b o t h z e r o p r i n t e r r o r , r e t u r n 0
i f ( ( a == 0 ) and ( b == 0 ) ) {
c e r r << "WARNING: gcd c a l l e d with two z e r o s . \ n " ;
return 0 ;
}
 MAT 685:
C++ for Math-
ematicians
C++ Implementation
John Perry

Background
Listing 5: add to gcd_recursive.cpp (p. 2)
Implementa-
i f ( b == 0) {
tion x = 1; y = 0;
Passing by reference return a;
Overloading
}
Implementation
i f ( a == 0) {
Summary x = 0; y = 1;
return b;
}

long c = a % b ;
long q = a / b ;

long d = gcd ( b , c , x , y ) ;
long new_x = y ;
long new_y = x − q ∗y ;
x = new_x ;
y = new_y ;
i f ( a _ n e g ) x = −x ;
i f ( b_neg ) y = −y ;
return d ;
}
 MAT 685:
C++ for Math-
ematicians
A basic test program
John Perry

Background

Implementa-
tion Listing 6: test_xgcd.cpp (place in same folder)
Passing by reference
Overloading # i n c l u d e <i o s t r e a m >
Implementation
using s t d : : c i n ; using s t d : : cout ; using s t d : : endl ;
Summary # i n c l u d e " gcd . hpp "

i n t main ( ) {
long a , b , x , y ;
c o u t << " E n t e r two numbers : " << e n d l ;
c i n >> a >> b ;
c o u t << " gc d ( " << a << " , " << b << " ) = " ;
c o u t << gcd ( a , b , x , y ) << " = " ;
c o u t << a << ’ ∗ ’ << x ;
c o u t << " + " << b << ’ ∗ ’ << y << e n d l ;
}
 MAT 685:
C++ for Math-
ematicians
Building from command line,
John Perry
running
Background

Implementa-
tion
Passing by reference $ c l a n g++ −c g c d _ r e c u r s i v e . cpp
Overloading $ c l a n g++ −o t e s t _ g c d g c d _ r e c u r s i v e . o t e s t _ x g c d . cpp
Implementation

Summary
 MAT 685:
C++ for Math-
ematicians
Building from command line,
John Perry
running
Background

Implementa-
tion
Passing by reference $ c l a n g++ −c g c d _ r e c u r s i v e . cpp
Overloading $ c l a n g++ −o t e s t _ g c d g c d _ r e c u r s i v e . o t e s t _ x g c d . cpp
Implementation

Summary
Usage:
$ ./test_xgcd
Enter two numbers:
26 14
gcd(26,14) = 2 = 26*-1 + 14*2
$ ./test_xgcd
Enter two numbers:
26 -14
gcd(26,-14) = 2 = 26*-1 + -14*-2
 MAT 685:
C++ for Math-
ematicians
Careful with references, however!
John Perry
If we change the lines. . .
Background
cout << gcd(a,b,x,y) << " = ";
Implementa-
tion cout << a << ’*’ << x;
Passing by reference
Overloading
Implementation . . . to. . .
Summary
cout << gcd(a,b,x,y) << " = " << a
<< ’*’ << x;

. . . we get a slightly wrong result!

$ ./test_xgcd
Enter two numbers:
26 14
gcd(26,14) = 2 = 26*0 + 14*2

Why?
 MAT 685:
C++ for Math-
ematicians
Careful with references, however!
John Perry
If we change the lines. . .
Background
cout << gcd(a,b,x,y) << " = ";
Implementa-
tion cout << a << ’*’ << x;
Passing by reference
Overloading
Implementation . . . to. . .
Summary
cout << gcd(a,b,x,y) << " = " << a
<< ’*’ << x;

. . . we get a slightly wrong result!

$ ./test_xgcd
Enter two numbers:
26 14
gcd(26,14) = 2 = 26*0 + 14*2

Why?

Answer
 MAT 685:
C++ for Math-
ematicians
Outline
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
1 Background
Implementation

Summary

2 Implementation
Passing by reference
Overloading
Implementation

3 Summary
 MAT 685:
C++ for Math-
ematicians
Summary
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Implementation

Summary • Math stuff

• Extended Euclidean Algorithm
• Programming stuff
• pass by reference
• overloading
 MAT 685:
C++ for Math-
ematicians
Homework
John Perry

Background

Implementa-
tion
Passing by reference
Overloading
Implementation

Summary

pp. 49-50 #3.1, 3.7, 3.9