pow() implementation in cmath and efficient replacement [duplicate] - c++

This question already has answers here:
The most efficient way to implement an integer based power function pow(int, int)
(21 answers)
Closed 8 years ago.
I've read that cmath calculates pow(a,b) by performing exp(b*log(a)). This should not be used when b is an integer, since it slows down calculations a lot. What alternatives are there when
calculating a lot of successive pow()s with the same constant a
it is known beforehand that b will definitely be an integer?
I am looking for fast alternatives which are efficient in these particular scenarios.

There are a number of faster alternatives I've collected over the years that typically rely on a recursive implementation of the function, and bit shifts to handle multiplication when warranted. The following provide functions tailored to integer, float and double. They come with the normal disclaimer: while faster not all possible test have been run and the user should validate input is sane before calling and on return... blah, blah, blah.. But, they are pretty darn useful:
I believe proper attribution goes to Geeks for Geeks Pow(x,n) as pointed out by blue moon. I had long since lost the links.. That looks like them. (minus a tweak or two).
/* Function to calculate x raised to the power y
Time Complexity: O(n)
Space Complexity: O(1)
Algorithmic Paradigm: Divide and conquer.
*/
int power1 (int x, unsigned int y)
{
if (y == 0)
return 1;
else if ((y % 2) == 0)
return power1 (x, y / 2) * power1 (x, y / 2);
else
return x * power1 (x, y / 2) * power1 (x, y / 2);
}
/* Function to calculate x raised to the power y in O(logn)
Time Complexity of optimized solution: O(logn)
*/
int power2 (int x, unsigned int y)
{
int temp;
if (y == 0)
return 1;
temp = power2 (x, y / 2);
if ((y % 2) == 0)
return temp * temp;
else
return x * temp * temp;
}
/* Extended version of power function that can work
for float x and negative y
*/
float powerf (float x, int y)
{
float temp;
if (y == 0)
return 1;
temp = powerf (x, y / 2);
if ((y % 2) == 0) {
return temp * temp;
} else {
if (y > 0)
return x * temp * temp;
else
return (temp * temp) / x;
}
}
/* Extended version of power function that can work
for double x and negative y
*/
double powerd (double x, int y)
{
double temp;
if (y == 0)
return 1;
temp = powerd (x, y / 2);
if ((y % 2) == 0) {
return temp * temp;
} else {
if (y > 0)
return x * temp * temp;
else
return (temp * temp) / x;
}
}

Non-recursive non-floating point answer
Replace uintmax_t/intmax_t with the type of your desire. Overflow not detected.
uintmax_t powjuu(unsigned x, unsigned y) {
uintmax_t z = 1;
uintmax_t base = x;
while (y) {
if (y & 1) { // or y%2
z *= base;
}
y >>= 1; // or y /= 2
base *= base;
}
return z;
}
intmax_t powjii(int x, int y) {
if (y < 0) {
switch (x) {
case 0:
return INTMAX_MAX;
case 1:
return 1;
case -1:
return y % 2 ? -1 : 1;
}
return 0;
}
intmax_t z = 1;
intmax_t base = x;
while (y) {
if (y & 1) {
z *= base;
}
y >>= 1;
base *= base;
}
return z;
}

You may want to check this. It's a fast algorithm to replace the pow function.

Related

Want to minimize my code so it consumes time less than 1 sec. It uses concept of modular exponentiation .Correct output but exceeding time limit

The below code is to calculate 2^n where n is equal to 1 <= n <= 10^5. So to calculate such large numbers I have used concept of modular exponentian. The code is giving correct output but due to large number of test cases it is exceeding the time limit. I am not getting a way to minimize the solution so it consumes less time. As the "algo" function is called as many times as the number of test cases. So I want to put the logic used in "algo" function in the main() function so it consumes time less than 1 sec and also gives the correct output. Here "t" represents number of test cases and it's value is 1 <= t <= 10^5.
Any suggestions from your side would be of great help!!
#include<iostream>
#include<math.h>
using namespace std;
int algo(int x, int y){
long m = 1000000007;
if(y == 0){
return 1;
}
int k = algo(x,y/2);
if (y % 2 == 1){
return ((((1ll * k * k) % m) * x) % m);
} else if (y % 2 == 0){
return ((1ll * k * k) % m);
}
}
int main(void)
{
int n, t, k;
cin>>t; //t = number of test cases
for ( k = 0; k < t; k++)
{
cin >> n; //power of 2
cout<<"the value after algo is: "<<algo(2,n)<<endl;
}
return 0;
}
You can make use of binary shifts to find powers of two
#include <iostream>
using namespace std;
int main()
{
unsigned long long u = 1, w = 2, n = 10, p = 1000000007, r;
//n -> power of two
while (n != 0)
{
if ((n & 0x1) != 0)
u = (u * w) % p;
if ((n >>= 1) != 0)
w = (w * w) % p;
}
r = (unsigned long)u;
cout << r;
return 0;
}
This is the function that I often use to calculate
Any integer X raised to power Y modulo M
C++ Function to calculate (X^Y) mod M
int power(int x, int y, const int mod = 1e9+7)
{
int result = 1;
x = x % mod;
if (x == 0)
return 0;
while (y > 0)
{
if (y & 1)
result = ( (result % mod) * (x % mod) ) % mod;
y = y >> 1; // y = y / 2
x = ( (x % mod) * (x % mod) ) % mod;
}
return result;
}
Remove the Mod if you don't want.
Time Complexity of this Function is O(log2(Y))
There can be a case of over flow so use int , long , long long etc as per your need.
Well your variables won't sustain the boundary test cases, introducing 2^10000, 1 <= n <= 10^5. RIP algorithms
19950631168807583848837421626835850838234968318861924548520089498529438830221946631919961684036194597899331129423209124271556491349413781117593785932096323957855730046793794526765246551266059895520550086918193311542508608460618104685509074866089624888090489894838009253941633257850621568309473902556912388065225096643874441046759871626985453222868538161694315775629640762836880760732228535091641476183956381458969463899410840960536267821064621427333394036525565649530603142680234969400335934316651459297773279665775606172582031407994198179607378245683762280037302885487251900834464581454650557929601414833921615734588139257095379769119277800826957735674444123062018757836325502728323789270710373802866393031428133241401624195671690574061419654342324638801248856147305207431992259611796250130992860241708340807605932320161268492288496255841312844061536738951487114256315111089745514203313820202931640957596464756010405845841566072044962867016515061920631004186422275908670900574606417856951911456055068251250406007519842261898059237118054444788072906395242548339221982707404473162376760846613033778706039803413197133493654622700563169937455508241780972810983291314403571877524768509857276937926433221599399876886660808368837838027643282775172273657572744784112294389733810861607423253291974813120197604178281965697475898164531258434135959862784130128185406283476649088690521047580882615823961985770122407044330583075869039319604603404973156583208672105913300903752823415539745394397715257455290510212310947321610753474825740775273986348298498340756937955646638621874569499279016572103701364433135817214311791398222983845847334440270964182851005072927748364550578634501100852987812389473928699540834346158807043959118985815145779177143619698728131459483783202081474982171858011389071228250905826817436220577475921417653715687725614904582904992461028630081535583308130101987675856234343538955409175623400844887526162643568648833519463720377293240094456246923254350400678027273837755376406726898636241037491410966718557050759098100246789880178271925953381282421954028302759408448955014676668389697996886241636313376393903373455801407636741877711055384225739499110186468219696581651485130494222369947714763069155468217682876200362777257723781365331611196811280792669481887201298643660768551639860534602297871557517947385246369446923087894265948217008051120322365496288169035739121368338393591756418733850510970271613915439590991598154654417336311656936031122249937969999226781732358023111862644575299135758175008199839236284615249881088960232244362173771618086357015468484058622329792853875623486556440536962622018963571028812361567512543338303270029097668650568557157505516727518899194129711337690149916181315171544007728650573189557450920330185304847113818315407324053319038462084036421763703911550639789000742853672196280903477974533320468368795868580237952218629120080742819551317948157624448298518461509704888027274721574688131594750409732115080498190455803416826949787141316063210686391511681774304792596709376
Fear not my friend, someone did tried to solve the problem https://www.quora.com/What-is-2-raised-to-the-power-of-50-000, you are looking for Piyush Michael's answer , here is his sample code
#include <stdio.h>
int main()
{
int ul=16,000;
int rs=50,000;
int s=0,carry[ul],i,j,k,ar[ul];
ar[0]=2;
for(i=1;i<ul;i++)ar[i]=0;
for(j=1;j<rs;j++)
{for(k=0;k<ul;k++)carry[k]=0;
for(i=0;i<ul;i++)
{ar[i]=ar[i]*2+carry[i];
if(ar[i]>9)
{carry[i+1]=ar[i]/10;
ar[i]=ar[i]%10;
}
}
}
for(j=ul-1;j>=0;j--)printf("%d",ar[j]);
for(i=0;i<ul-1;i++)s+=ar[i];
printf("\n\n%d",s);
}

Multiplying using recursive function C++

I did a recursive function to calculate x*y with x and y are all integers (x and y >= 0). My formula is:
x * y =
0, if x is equal 0
(x >> 1)*(y << 1), if x is an even number
(x >> 1)*(y << 1) + y, if x is an odd number
"<<" and ">>" are Left Shift and Right Shift Bitwise Operator. Here is my code:
int multiply(int x, int y) {
int y1 = 0;
if (x == 0) return 0;
else if (x % 3 == 0) {
y1 = y;
x = x >> 1;
y = y << 1;
return (multiply(x, y) + y1);
}
else if (x % 2 == 0) {
x = x >> 1;
y = y << 1;
return multiply(x, y);
}
}
The recursive function above is supposed to return (x*y) value but they were all wrong when i tested and i don't know why. What did i do wrong? How can i fix this?
Your problem is wit x % 3, what happens if x = 5? you skip it. Here is improved version of your code.
int multiply(int x, int y) {
if (x == 0)
return 0;
else if (x % 2 == 1)
return (multiply(x >> 1, y << 1) + y);
return multiply(x >> 1, y << 1);
}
or maybe even this:
int multiply(int x, int y) {
if (x == 0)
return 0;
int m = multiply(x >> 1, y << 1);
if (x % 2 == 1)
m += y;
return m;
}
Here is super fast version suggested by Andy:
int multiply(int x, int y) {
if (x == 0)
return 0;
int m = multiply(x >> 1, y << 1);
if (x & 1)
m += y;
return m;
}
As a challenge of speed, here is non recursive version:
int multiply (int x, int y) {
int y1 = 0;
for (; x > 0; x = (x >> 1), y = (y << 1))
if (x&1)
y1 += y;
return y1;
}
NOTE: I know this question is about recursive method but just as a challenge I wrote non-recursive algorithm.
You are not checking if x is odd correctly here:
else if (x % 3 == 0) { // e.g. fails on x = 1
Instead, you need to do
else if (x % 2 == 1) {
Here's a demo.
Note that this makes the following else check for even values of x redundant:
else if (x % 2 == 0) { // can just be an unconditional else
Also, since you are returning from the x == 0, and x % 2 == 1 branches, the else conditions can be removed as well. You can also factor out the repeated code to make the function simpler, like this:
int multiply(int x, int y) {
if (x == 0) return 0;
if (x % 2 == 1)
return (multiply(x >> 1, y << 1) + y);
else
return multiply(x >> 1, y << 1);
}
Here's a demo.
This is what i feel is the simplest approach to carry out recursive multiplication.
Do let me know if its efficient enough for you.
#include<iostream>
using namespace std;
float multiply(float a, float b) {
//no zeros bro
if (b == 0)
return 0;
//let the recursion begin
if (b > 0)
return x + multiply(a, b - 1);
//negatives stay away pliz
if (y < 0)
return -multiply(a, -b);
}
int main() {
float a, b, result;
cout << "Enter the two numbers";
cin >> a >> b;
result = multiply(a, b);
//And the result is.................
cout << result;
return 0;
}
Recursive function related to multiplication of natural numbers:
int multCool(int a, int b)
{
return (b==1 ? a: a+ multCool(a,b-1));
}

Using brute-force for solving equations

I am trying to solve a very simple equation in c++ by brute force. The basic idea is to run up or down the value of x until the left side of the equation is equal to the right side. I am not getting any errors, but the value of x is always either 0.01 or -0.01. I assume that my do/while loop is flawed, but you are all probably more experienced than I am, so any help is appreciated.
#include <iostream>
using namespace std;
int main()
{
double x, y, z, q, w;
x = 0;
cout << "enter y*x + q = z*x + w in order of appearance" << endl;
cin >> y;
cin >> q;
cin >> z;
cin >> w;
if ((y-z)/(w-q) > 0) // checks if x is positive
{
do
{
(x = x + 0.01);
} while ((y * x + q) == (z * x + w));
}
else if ((y - z) / (w - q) < 0) // checks if x is negative
{
do
{
(x = x - 0.01);
} while ((y * x + q) == (z * x + w));
}
else
{
x = 0;
}
cout << "x is " << x << endl;
return 0;
}
Thanks!
A few things.
First, when comparing floats, you probably want to do so within a tight range, frequently referred to as epsilon. This is especially true because you're incrementing by a fairly wide margin -- 0.01. You are probably skipping over the value you want.
What I would do is pay attention to:
Am I getting closer or further from the answer?
And did I skip over the answer?
Some code:
float leftSide = y * x + q;
float rightSide = z * x + w;
float delta = leftSide - rightSide;
if (abs(delta) < epsilon) {
// You're really close
}
Note also this will never work if y and z are identical values unless q and w are as well.

How do I fix this bug with arctanx function in c++

I've recently been given a problem by my teacher about some mathematical equation / formula called the arctanx formula. The question is:
According to the Arctanx(x) = x - ((x ^ 3) / 3) + ((x ^ 5) / 5) - ((x ^
7) / 7) + ...and π = 6 * arctanx(1 / sqrt(3)), Create function arctanx(x)
, and find pi when the last "number"(like this ((x ^ y) / y)) is right before
and bigger than 10 ^ -6, or you can say that no "number" can be smaller than
that number without being smaller than 10 ^ -6.
I tried to code it out, but there is a bug in it.
# include<iostream>
# include<math.h>
using namespace std;
float arctanx() {
long double pi = 3.1415926535897;
int i = 0; // 0 = +, 1 = -
float sum = 0;
float lsum;
for (int y = 1; y < pi; y += 2) {
if (lsum > 0.000001) {
if (i == 0) {
lsum = pow(1 / sqrt(3), y) / y;
sum += pow(1 / sqrt(3), y) / y;
i++;
} else if (i == 1) {
lsum = pow(1 / sqrt(3), y) / y;
sum -= pow(1 / sqrt(3), y) / y;
i--;
}
} else {
break;
}
}
sum = sum * 6;
return sum;
}
int main() {
cout << arctanx();
return 0;
}
It should have a output of some number not equal to zero, but I got 0 from running this.
Your program has Undefined Behavior because you are using the uninitialized float lsum; in the comparison if (lsum > 0.000001).
What probably happens in your case is that lsum happens to be less than or equal to 0.000001 and your for immediately breaks without doing anything causing your function to return 0 * 6 which is obviously 0.
Create function arctanx(x)
The function defined in the posted code doesn't accept any parameter, it just uses the hardwired (and repeated) value 1 / sqrt(3) and tries to return an approximated value of π instead of the arctangent of x.
It also has undefined behavior, beeing lsum uninitialized (therefore having an indeterminate value) when it is first used in the comparison inside the loop.
Consider this implementation, but be advised that this particular polinomial expansion diverges for values of x greater than 1.
#include <iostream>
#include <iomanip>
#include <cmath>
double arctanx(double x);
int main()
{
double pi = 6.0 * arctanx(1.0 / std::sqrt(3));
std::cout << std::setprecision(8) << pi << '\n';
}
double arctanx(double x)
{
// You can take advantage of a running power, instad of calculating
// pow(x, i) at every iteration
double sq_x = x * x;
double pow_x = x * sq_x;
double err = 1e-6;
// Instead of keeping track of the alternating sign, you can use
// two separate partial sums
double sum_pos_term = x;
double sum_neg_term = 0.0;
for (int i = 3; i < 33; i += 2) // <- Limit the number of iterations
{
if (pow_x < err * i)
break;
sum_neg_term += pow_x / i;
i += 2;
pow_x *= sq_x;
if (pow_x < err * i)
break;
sum_pos_term += pow_x / i;
pow_x *= sq_x;
}
return sum_pos_term - sum_neg_term;
}

Square Root in C/C++

I am trying to implement my own square root function which gives square root's integral part only e.g. square root of 3 = 1.
I saw the method here and tried to implement the method
int mySqrt(int x)
{
int n = x;
x = pow(2, ceil(log(n) / log(2)) / 2);
int y=0;
while (y < x)
{
y = (x + n / x) / 2;
x = y;
}
return x;
}
The above method fails for input 8. Also, I don't get why it should work.
Also, I tried the method here
int mySqrt(int x)
{
if (x == 0) return 0;
int x0 = pow(2, (log(x) / log(2))/2) ;
int y = x0;
int diff = 10;
while (diff>0)
{
x0 = (x0 + x / x0) / 2; diff = y - x0;
y = x0;
if (diff<0) diff = diff * (-1);
}
return x0;
}
In this second way, for input 3 the loop continues ... indefinitely (x0 toggles between 1 and 2).
I am aware that both are essentially versions of Netwon's method but I can't figure out why they fail in certain cases and how could I make them work for all cases. I guess i have the correct logic in implementation. I debugged my code but still I can't find a way to make it work.
This one works for me:
uintmax_t zsqrt(uintmax_t x)
{
if(x==0) return 0;
uintmax_t yn = x; // The 'next' estimate
uintmax_t y = 0; // The result
uintmax_t yp; // The previous estimate
do{
yp = y;
y = yn;
yn = (y + x/y) >> 1; // Newton step
}while(yn ^ yp); // (yn != yp) shortcut for dumb compilers
return y;
}
returns floor(sqrt(x))
Instead of testing for 0 with a single estimate, test with 2 estimates.
When I was writing this, I noticed the result estimate would sometimes oscillate. This is because, if the exact result is a fraction, the algorithm could only jump between the two nearest values. So, terminating when the next estimate is the same as the previous will prevent an infinite loop.
Try this
int n,i;//n is the input number
i=0;
while(i<=n)
{
if((i*i)==n)
{
cout<<"The number has exact root : "<<i<<endl;
}
else if((i*i)>n)
{
cout<<"The integer part is "<<(i-1)<<endl;
}
i++;
}
Hope this helps.
You can try there C sqrt implementations :
// return the number that was multiplied by itself to reach N.
unsigned square_root_1(const unsigned num) {
unsigned a, b, c, d;
for (b = a = num, c = 1; a >>= 1; ++c);
for (c = 1 << (c & -2); c; c >>= 2) {
d = a + c;
a >>= 1;
if (b >= d)
b -= d, a += c;
}
return a;
}
// return the number that was multiplied by itself to reach N.
unsigned square_root_2(unsigned n){
unsigned a = n > 0, b;
if (n > 3)
for (a = n >> 1, b = (a + n / a) >> 1; b < a; a = b, b = (a + n / a) >> 1);
return a ;
}
Example of usage :
#include <assert.h>
int main(void){
unsigned num, res ;
num = 1847902954, res = square_root_1(num), assert(res == 42987);
num = 2, res = square_root_2(num), assert(res == 1);
num = 0, res = square_root_2(num), assert(res == 0);
}
Source