改进快速排序

问题描述

如果可能的话，我该如何改善以下快速排序(性能明智).有什么建议?

If possible, how can I improve the following quick sort(performance wise). Any suggestions?

void main()
    {
      quick(a,0,n-1);
    }

    void quick(int a[],int lower,int upper)
    {
       int loc;
       if(lower<upper)
       {
        loc=partition(a,lower,upper);
        quick(a,lower,loc-1);
        quick(a,loc+1,upper);

       }
    }

    /* Return type: int
      Parameters passed: Unsorted array and its lower and upper bounds */

    int partition(int a[],int lower,int upper)
    {
      int pivot,i,j,temp;
      pivot=a[lower];
      i=lower+1;
      j=upper;
      while(i<j)
        {
            while((i<upper)&&(a[i]<=pivot))
            i++;
            while((a[j]>pivot))
            j--;
            if(i<j)
                {
                    temp=a[i];
                    a[i]=a[j];
                    a[j]=temp;
                }

        }//end while

        if(pivot>a[j])
        {
             temp=a[j];
             a[j]=a[lower];
             a[lower]=temp;
        }

         return(j);

}//end partition

推荐答案

快速排序算法易于并行化.如果您有多个内核可以使用，则可以看到相当多的速度提升.根据您的数据集的大小，它可以轻松地为您提供比任何其他优化都更高的速度.但是，如果您只有一个处理器或相对较小的数据集，那么速度不会有太大提高.

The quicksort algorithm is easily parallelized. If you have multiple cores to work with, you could see quite a bit of speed up. Depending on how large your data set is, it could easily provide you with more speed up than any other optimization. However, if you have only one processor or a relatively small data set, there won't be much of a speed up.

如果可能的话，我可以详细说明.

I could elaborate more if this is a possibility.

这篇关于改进快速排序的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！