Unable to calculate factorial , getting output as 0 in C

Question

Tried calculating factorial of 65, got correct output. Anything greater than 65 results in output of 0. Shocking since I'm using unsigned long int. What is amiss ?

Code:

#include <stdio.h>

void factorial(int unsigned long);
int main()
{
    int unsigned long num, result;
    printf("\nEnter number to obtain factorial : ");
    scanf("%ld", &num);
    factorial(num);
}
void factorial (int unsigned long x)
{
    register int unsigned long f = 1;
    register int unsigned long i;
    for (i=x;i>=1;i--)
        f= f*i;
    printf("\nFactorial of %lu = %lu\n",x,f);
}

While C **allows** you to write `int unsigned long`, a seasoned programmer never writes that, using instead `unsigned long`, or `unsigned long int`. — Antti Haapala -- Слава Україні, Jul 04 '20 at 08:28
a tip about naming: when using single letter variables, avoid variables that may suggest the wrong type name. For example, the name `f` might lead you to think of the variable as a `float`. As your code becomes longer or more complex, these things become more important. — Myst, Jul 04 '20 at 08:31
How do you know you got the correct output when you calculate 65! ? — M. Nejat Aydin, Jul 04 '20 at 08:35
@M.NejatAydin Your right, my value for factorial 65 generated by my code is 9223372036854775808 , very off from the real value. — user13863346, Jul 04 '20 at 09:12

score 2 · Accepted Answer · edited Jul 04 '20 at 08:27

2

You certainly did not get the correct result for 65! log2(65!) is just over 302 bits (Google it), so you'd need a long int of at least 303 bits to calculate that correctly. There is no computer in the world where long int is over 300 bits (let's see how this answer ages!).

The largest factorial you can compute in 64 bits is 20! (which is about 2.4e18).

edited Jul 04 '20 at 08:27

SomeWittyUsername

18,025
3
42
85

answered Jul 04 '20 at 08:16

John Zwinck

239,568
38
324
436

Well then , how can Google / any other online calculator give the value of factorial of any number above 20 ? Does is break down the 300+ bits into chunks of 64 , like 64 bit is broken down for 32-bit CPUs ? – user13863346 Jul 04 '20 at 09:05
@user13863346: They use https://en.wikipedia.org/wiki/Arbitrary-precision_arithmetic – John Zwinck Jul 04 '20 at 09:10
Any way a noob like me could try implementing something similar ? – user13863346 Jul 04 '20 at 09:31
Sure, you can use the GMP library: https://gmplib.org/ – John Zwinck Jul 04 '20 at 10:32

Sai Sreenivas · Answer 2 · 2020-07-04T08:35:03.750

0

Adding on to @JohnZwinck, the maximum value for a variable of type unsigned long long (ULLONG_MAX) is 18446744073709551615. So all the values greater than 20! are going to have a garbage value

You can refer to this for more information.

edited Jul 04 '20 at 08:35

answered Jul 04 '20 at 08:20

Sai Sreenivas

1,690
1
7
16

1

That's not the maximum value, it is the *minimal maximum value*. – Antti Haapala -- Слава Україні Jul 04 '20 at 08:29
This does not answer to the question. – M. Nejat Aydin Jul 04 '20 at 08:33

Unable to calculate factorial , getting output as 0 in C

2 Answers2