are srand() and rand() 's implementations standardized?

Hi,
Are the implementations of srand() and rand() standardized? eg. If I use the same seed, can I expect to get the same sequence from gcc and Microsoft compiler? If not, can someone please suggest a good general-purpose (not for cryptography) pseudo-random generator?

Thanks

Originally Posted by cyberfish

Hi,
Are the implementations of srand() and rand() standardized? eg. If I use the same seed, can I expect to get the same sequence from gcc and Microsoft compiler? If not, can someone please suggest a good general-purpose (not for cryptography) pseudo-random generator?

You are not guaranteed to get the same results. As for an algorithm I recommend Mersenne Twister. Search for "mt19937.c"

No.

On my Vista machine:

Code:

C:\Users\...\Documents\Code>randtest
4404, 25954, 17763, 29679, 4113, 18955, 8176, 17911, 814, 2645,

On a FreeBSD machine I have access to:

Code:

$ gcc -o randtest randtest.c
$ ./randtest
22470959, 1859769688, 524664131, 462195135, 665282746, 1607245740, 1929840214, 1378956057, 470931575, 1469741830,
$ cat randtest.c
#include <stdio.h>
#include <stdlib.h>

int main()
{
        int x;

        srand(1337);
        for(x = 0; x < 10; ++x)
        {
                printf("&#37;u, ", rand());
        }
        printf("\n");

        return 0;
}
$

Thank you for your help. I will certainly look into Mersenne Twister.

The TR1 libraries contain the random library which contains a framework for random number generation, including a Mersenne Twister.
If your compiler doesn't provide a TR1 implementation yet, you can get Boost's version.

http://boost.org/

Incidentally, Prelude also has an implementation of the Mersenne Twister algorithm at her eternally confuzzling website. The download link is under "Libraries".

I see... Thanks for your replies.

I guess I will just use mt19937.c for the time being =).