First, the function advances the internal state by one using a transition algorithm that produces a twist on the selected element in the state as if this was x[i] in the following piece of code:
1 2 3
UIntType Mask = (1u<<w)-1, UMask= (Mask << r) & Mask, LMask = (~UMask) & Mask,
Y = (x[i]&UMask)|(x[(i+1)%n]&LMask); // upper bits of x[i] and lower of x[i+1]
x[i]=x[(i+m)%n]^(Y>>1)^((Y&1)*a); // mersenne twister linear transformation
With each of the values being their corresponding class template parameters.
To improve the uniformity of the random value produced, the function uses a generation algorithm to return a tempered version of the selected element in the state sequence: it applies the same transformation applied on y in the following piece of code:
1 2 3 4
y^=(y>>u)&d;
y^=(y<<s)&b;
y^=(y<<t)&c;
y^=y>>l;
With each of the values being again their corresponding class template parameters.
Note: not fully tested, plese report any inaccuracies.
Parameters
None
Return value
A new random number. result_type is a member type, defined as an alias of the first class template parameter (UIntType).
Example
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
// mersenne_twister_engine::operator()
#include <iostream>
#include <chrono>
#include <random>
int main ()
{
// obtain a seed from the system clock:
unsigned seed = std::chrono::system_clock::now().time_since_epoch().count();
std::mt19937 generator (seed); // mt19937 is a standard mersenne_twister_engine
std::cout << "Random value: " << generator() << std::endl;
return 0;
}