Не работает при 3, 6, 7 и т.д. потоках (segfault)

#include <omp.h>
#include <malloc.h>
#include <stdio.h>
#include <xmmintrin.h>

#define N 1000

int main() {
    double* vector1 = malloc(sizeof(double) * N);
    double* vector2 = malloc(sizeof(double) * N);

    for (unsigned int i = 0; i < N; i++) {
        vector1[i] = i;
        vector2[i] = 1 / (double) i;
    }

    #pragma omp parallel
    {
        const int thread_num = omp_get_thread_num();
        const int num_threads = omp_get_num_threads();

        const int k1 = thread_num * N / num_threads;
        const int k2 = (thread_num + 1) * N / num_threads;

        __m128d* x = (__m128d*) (vector1 + k1);
        __m128d* y = (__m128d*) (vector2 + k1);

        __m128d tmp1 = x[0];
        __m128d tmp2 = y[0];
    }

    free(vector1);
    free(vector2);

    return 0;
}