为什么这种并行二进制搜索要比顺序搜索花费更多时间？

Question

我在openMP c ++中具有以下代码，该代码顺序且并行地运行迭代二进制搜索。根据输出，为什么即使并行运行4个部分，并行搜索仍比顺序搜索花费更多时间？

#include <iostream>
#include <algorithm>
#include <omp.h>
using namespace std;

#define MAX1 500000

int binary_search(int arr[], int x, int low, int high)
{
    while(low <= high)
    {
        int mid = low + (high-low)/2;
        if(arr[mid] == x)
            return mid;
        if(arr[mid] < x)
            low = mid + 1;
        else
            high = mid - 1;
    }
    return -1;
}

void binary()
{
    int array[MAX1];
    int loc, loc1, loc2, loc3, loc4, k;
    omp_set_num_threads(4);

    for(int i=0; i<MAX1; ++i)
        array[i] = rand()%1000;

    sort(array, array+MAX1);

    cout<<"Enter number to search : ";
    cin>>k;
    int mid = MAX1/2;
    double start_time = omp_get_wtime();
    loc1 = binary_search(array, k, 0, mid);
    loc2 = binary_search(array, k, mid+1, MAX1-1);
    double end_time = omp_get_wtime()-start_time;
    cout<<loc1<<"\t"<<loc2<<endl;
    cout<<"Sequential Binary search runtime : "<<end_time<<endl;

    int one_quart = mid/2;
    int three_quart = 3*one_quart;
    start_time = omp_get_wtime();
    #pragma omp parallel sections
    {
        #pragma omp section
        loc1 = binary_search(array, k, 0, one_quart);
        #pragma omp section
        loc2 = binary_search(array, k, one_quart+1, mid);
        #pragma omp section
        loc3 = binary_search(array, k, mid+1, three_quart);
        #pragma omp section
        loc4 = binary_search(array, k, three_quart+1, MAX1-1);
    }
    end_time = omp_get_wtime()-start_time;
    cout<<loc1<<"\t"<<loc2<<"\t"<<loc3<<"\t"<<loc4<<endl;
    cout<<"Parallel Binary search runtime : "<<end_time<<endl;
}

输出：

输入要搜索的号码：777

-1 388671

顺序二进制搜索运行时：3.478e-06

-1 -1 -1 388671

并行二进制搜索运行时：0.000629687