#include <iostream>
#include <omp.h>

int main(int argc, const char **argv)
{
    #pragma omp parallel
    {
        int nloops = 0;

        #pragma omp for 
        for (int i=0; i<1000000; ++i)
        {
            ++nloops;
        }

        int thread_id = omp_get_thread_num();

        std::cout << "Thread " << thread_id << " performed "
                  << nloops << " iterations of the loop.\n";
    }

    return 0;
}

Mein Output:

Thread Thread Thread Thread 910 performed 833336Thread performed 83333 iterations of the loop.
Thread 0 performed 83334 iterations of the loop.
Thread 3 performed 83334 iterations of the loop.
iterations of the loop.
performed 83333 iterations of the loop.
Thread 2 performed 83334 iterations of the loop.
Thread 4 performed 83333 iterations of the loop.
11 performed 83333 iterations of the loop.
7 performed 83333 iterations of the loop.
Thread 1 performed 83334 iterations of the loop.
Thread 8 performed 83333 iterations of the loop.
Thread 5 performed 83333 iterations of the loop.

Ich verstehe nicht so ganz, warum das nicht funktioniert.

Ich wollte mir mal OpenMP anschauen

OpenMP ist nicht mehr zeitgemäss. Nimm lieber TBB oder so.