Random Number Generator is always generating the same numbers

Question

I am trying to fill my array with random numbers using the following piece of code

#include<iostream>
#include<random>

int main(){
int n = 5;
int list[10];
std::random_device rd;
std::mt19937 eng(rd());
std::uniform_int_distribution<> distr(0, 1000);

for(int i=0;i<n;i++)
    list[i] = distr(eng);

std::cout<<"The list of elements is: ";
for(int i=0;i<n;i++)
    std::cout<<list[i]<<" ";
}

For n = 5, I always get the same output

562 726 348 916 6

For n = 6, I always get the same output

562 726 348 916 6 594

These numbers arent random, I also checked the entropy

std:cout<<rd.entropy();

This gives me the output

What am I doing wrong and how do I get random numbers in my array?

You might want to include what compiler and version you are using and what platform you are on. Your code works for me in VS2013. — The Dark, Apr 12 '16 at 07:16
Possible duplicate of [C++, twenty numbers not random?](http://stackoverflow.com/questions/2159490/c-twenty-numbers-not-random) — , Apr 12 '16 at 07:30
@JarrodRoberson, that post is about `srand`, but this one is about `random_device`. The latter should produce better randomness. — Serge Rogatch, Apr 12 '16 at 07:32
I had the same problem with g++ libraries in Windows. You can use functions to seed the deterministic generator like [here](http://stackoverflow.com/a/34682790/4944425). — Bob__, Apr 12 '16 at 07:53
@FirstStep: No, `srand()` is the seed for `rand()`, which is obsolete. — MSalters, Apr 12 '16 at 11:36
@Msalters I didn't understand. Could you please explain more? — Khalil Khalaf, Apr 12 '16 at 13:20

score 3 · Answer 1 · answered Apr 12 '16 at 10:54

Call the entropy() member function on std::random_device to find out whether your implementation implements it properly:

std::random_device may be implemented in terms of an implementation-defined pseudo-random number engine if a non-deterministic source (e.g. a hardware device) is not available to the implementation. In this case each std::random_device object may generate the same number sequence.

(Source)

If this is the case, a call to entropy() will return 0:

A deterministic random number generator (e.g. a pseudo-random engine) has entropy zero.

If that is the case, you need to use a different fallback mechanism for seeding. For instance, you could use a time-based seed like in the old C-days.

On desktop platforms in particular, you should expect std::random_device to be implemented as a proper non-deterministic source though. If this is not the case, you might just be using a very old version of the standard library implementation. If you have the feeling that the implementation should support non-deterministic std::random_device but it does not, consider filing a bug report with your standard library maintainer.

score 1 · Answer 2 · answered Apr 12 '16 at 07:19

If a random hardware device is not available to the implementation, a pseudo random number engine is used. Is a random number hardware device available? By that, I mean not only physically there, but also available to that particular implementation of std::random_device.

Random Number Generator is always generating the same numbers

2 Answers2