多线程快速排序比java中的普通快速排序慢?

jq6vz3qz  于 2021-07-03  发布在  Java
关注(0)|答案(1)|浏览(400)

我比较了快速排序(普通,递归)和线程快速排序(使用线程)。我发现普通快速排序比线程快速排序快。是这样吗?我觉得我做错了什么。
代码如下:

public class QuickSort implements SortAlgorithm {
    public void Sort(Integer[] array) {
        Sort(array, 0, array.length-1);
    }

    private void Sort(Integer[] array, Integer low, Integer high) {
        if (low < high) {
            Integer pivot = partition(array, low, high);
            Sort(array, low, pivot-1);
            Sort(array, pivot+1, high);
        }
    }

    private Integer partition(Integer[] array, Integer low, Integer high) {
        Integer pivot = array[high];  
        Integer i = (low-1);
        for (Integer j=low; j<high; j++) 
            if (array[j] < pivot) { 
                i++; 
                Integer temp = array[i]; 
                array[i] = array[j]; 
                array[j] = temp; 
            } 

        Integer temp = array[i+1]; 
        array[i+1] = array[high]; 
        array[high] = temp; 

        return i+1; 
    }
}
public class ThreadsQuickSort extends Thread implements SortAlgorithm {
    private Integer low, high;
    private Integer[] array;

    public ThreadsQuickSort() {
        ;
    }

    public ThreadsQuickSort(Integer[] array, Integer low, Integer high) {
        this.array = array;
        this.low = low;
        this.high = high;
    }

    public void run() {
        Sort(array, low, high);
    }

    public void Sort(Integer[] array) {
        Sort(array, 0, array.length - 1);
    }

    private void Sort(Integer[] array, Integer low, Integer high) {
        if (low < high) {
            Integer pivot = partition(array, low, high);
            ThreadsQuickSort lowQSort = new ThreadsQuickSort(array, low, pivot - 1);
            lowQSort.start();
            ThreadsQuickSort highQSort = new ThreadsQuickSort(array, pivot + 1, high);
            highQSort.start();
            try {
                lowQSort.join();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
            try {
                highQSort.join();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }

    private Integer partition(Integer[] array, Integer low, Integer high) {
        Integer pivot = array[high];  
        Integer i = (low-1);
        for (Integer j=low; j<high; j++) 
            if (array[j] < pivot) { 
                i++; 
                Integer temp = array[i]; 
                array[i] = array[j]; 
                array[j] = temp; 
            } 

        Integer temp = array[i+1]; 
        array[i+1] = array[high]; 
        array[high] = temp; 

        return i+1; 
    }
}

main函数使用这两种算法对10000个项目的数组进行排序,并获得它们的执行时间。获得以下输出:

Single array sorted in 11 ms
The array was successfully sorted
Threaded array sorted in 6378 ms
The array was successfully sorted

我的问题是,这是否是良好的编码和输出是预期的。这与线程化快速排序的实现无关。

62lalag4

62lalag41#

每个线程大约70个微创建和开关。
螺纹的总和应在20000-1左右,即20000*70us=140000us=1400ms=1.4s
因此,这接近预期。

相关问题