使用openMP并行优化时出现负优化

Question

代码如下：

 double startTime1 = omp_get_wtime();
#pragma omp parallel sections shared(img1,img2) num_threads(2)
    {
#pragma omp section
        {
            img1.convertTo(img1, CV_64FC1, 1.0 / 255, 0);
            printf("section 0,tid=%ld\n", omp_get_thread_num());
        }
#pragma omp section
        {
            img2.convertTo(img2, CV_64FC1, 1.0 / 255, 0);
            printf("section 1,tid=%ld\n", omp_get_thread_num());
        }
    }
    double endTime1 = omp_get_wtime();
    printf("Time1: %lf\n", endTime1 - startTime1);

    double startTime2 = omp_get_wtime();
    img1.convertTo(img1, CV_64FC1, 1.0 / 255, 0);
    img2.convertTo(img2, CV_64FC1, 1.0 / 255, 0);
    double endTime2 = omp_get_wtime();
    printf("Time2: %lf\n", endTime2 - startTime2);

结果如下：

求解？

Answer 1

可能是因为并行的时候你print了内容，而串行的时候没有print，所以导致时间长吧，毕竟本来时间都很短，print占用了很多时间也不算过分