Speed comparison between matlab and c++ random number generator

I'm searching for an efficient way to create random integer vectors in C++. I made some comparison with an equivalent matlab code and the difference in terms of speed was vast; for large vectors matlab needs only 3% the time of C++. Is there a way to speed up the process?

    //********C++********//
    int m = 200000;
    float density = 1.0e-5;
    int nnz = density *m*m;

    uint32_t seed;
    std::mt19937 eng;
    std::srand(time(nullptr));

    seed = std::rand();
    eng.seed(seed);

    std::uniform_int_distribution<int> dist{ 0, m };

    auto gen = [&dist, &eng]() {
        return dist(eng);
    };

    //elapsed time 256.511 msecs.
    std::vector<int> v1(nnz), v2(nnz);
    std::generate(std::begin(v1), std::end(v1), gen);
    std::generate(std::begin(v2), std::end(v2), gen);
    //*******************//

    %********MATLAB********%
    m=200000;
    density = 1.e-5;
    nnz = density*m*m;

    tic;
        % Elapsed time is 9.842 milliseconds.
        v1=randi([0 m],nnz,1);
        v2=randi([0 m],nnz,1);
    toc;
    %***********************%