bitonic_sort 串行（递归和for循环）和并行（cuda）两个版本的代码实现

串行版本源码如下：

#include "stdio.h"
#include "iostream"
using namespace std;


bool ASCENDING=true;
bool DESCENDING=false;
bool dir=true;
void bitonicSort(int lo,int n,bool dir);
void bitonicMerge(int lo,int n,bool dir);
void exchange(int i,int j);
void compare(int i,int j,bool dir);
//input array
int a[8]={5,4,2,1,6,3,8,7};
int b[8]={5,4,2,1,6,3,8,7};
int N=8;

void bitonicSort(int lo,int n,bool dir)
{
	if (n>1)
	{
		int m=n/2;
		bitonicSort(lo,m,ASCENDING);
		bitonicSort(lo+m,m,DESCENDING);
		bitonicMerge(lo,n,dir);
	}
}


void bitonicMerge(int lo,int n,bool dir)
{
	if (n>1)
	{
		int m=n/2;
		for (int i = lo;i < lo+m;i++)
		compare(i,i+m,dir);
		bitonicMerge(lo,m,dir);
		bitonicMerge(lo+m,m,dir);
	}
}

void compare(int i,int j,bool dir)
{
	if(dir == (a[i] > a[j]))
		exchange(i,j);
}
void exchange(int i,int j)
{
	int t=a[i];
	a[i]=a[j];
	a[j]=t;
}



void main()
{

	bitonicSort(0,8,ASCENDING);
	for (int i=0;i < 8; i++)
	{
		cout<
   
   
    
    <
    
    
     
     <= N ;k=2*k)
	{
		for (j=k>>1;j > 0;j = j>>1)
		{
			for (i = 0;i < N;i++)
			{
				int ixj=i^j;
				if((ixj > i))
				{
					if((i&k)==0 && b[i] > b[ixj])
					{
						tem=b[i];
						b[i]=b[ixj];
						b[ixj]=tem;
					}
					if((i&k)!=0 && b[i] < b[ixj])
					{
						tem=b[i];
						b[i]=b[ixj];
						b[ixj]=tem;
					}
				}


			}
		}
	}
	for (int i=0;i < 8; i++)
	{
		cout<
     
     
      
      <
      
      
     
     
    
    
   
   

并行（cuda）版本源码如下：

#include "bitonic_sort.cuh"
#include "common.cuh"



extern float sum;

__global__ void bitonic_sort_step(float *dev_values, int j, int k)
{
  unsigned int i, ixj; /* Sorting partners: i and ixj */
  i = threadIdx.x + blockDim.x * blockIdx.x;
  ixj = i^j;
 
  /* The threads with the lowest ids sort the array. */
  if ((ixj)>i) {
    if ((i&k)==0) {
      /* Sort ascending */
      if (dev_values[i]>dev_values[ixj]) {
        /* exchange(i,ixj); */
        float temp = dev_values[i];
        dev_values[i] = dev_values[ixj];
        dev_values[ixj] = temp;
      }
    }
    if ((i&k)!=0) {
      /* Sort descending */
      if (dev_values[i]
     
     
      <<="1)" minor (j="k">
      >1; j>0; j=j>>1) {
		  bitonic_sort_step<<
      
      
       
       >>(dev_values, j, k);
	  }
  }

  cudaEventRecord(stop, 0); 
  cudaEventSynchronize(stop); 
  float elapsedTime; 

  cudaEventElapsedTime(&elapsedTime, start, stop); 
  printf("kernel耗时：%f毫秒\n",elapsedTime);
  sum = sum + elapsedTime;



  cudaMemcpy(values, dev_values, size, cudaMemcpyDeviceToHost);
  cudaFree(dev_values);
}