mpi_all_to_all.c.txt

MPI_ALLTOALL是组内进程之间完全的消息交换其中每一个进程都相其它所有的进 程发送消息同时每一个进程都从其它所有的进程接收消息

#include "mpi.h"
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <errno.h>

void main( int argc, char **argv )
{
int rank, size;
int chunk = 2;
/*发送到一个进程的数据块的大小*/
int i,j;
int *sb;
int *rb;
//int status, gstatus;

MPI_Init(&argc,&argv);
MPI_Comm_rank(MPI_COMM_WORLD,&rank);
MPI_Comm_size(MPI_COMM_WORLD,&size);
sb = (int *)malloc(size*chunk*sizeof(int));/*申请发送缓冲区*/
if ( !sb ) {
perror( "can't allocate send buffer" );

MPI_Abort(MPI_COMM_WORLD,EXIT_FAILURE);
}
rb = (int *)malloc(size*chunk*sizeof(int));/*申请接收缓冲区*/
if ( !rb ) {
perror( "can't allocate recv buffer");
free(sb);
MPI_Abort(MPI_COMM_WORLD,EXIT_FAILURE);
}
for ( i=0 ; i < size ; i++ ) {
for ( j=0 ; j < chunk ; j++ ) {
sb[i*chunk+j] = rank + i*chunk+j;/*设置发送缓冲区的数据*/
printf("myid=%d,send to id=%d, data[%d]=%d\n",rank,i,j,sb[i*chunk+j]);
rb[i*chunk+j] = 0;/*将接收缓冲区清0*/
}
}
/* 执行MPI_Alltoall 调用*/
MPI_Alltoall(sb,chunk,MPI_INT,rb,chunk,MPI_INT,
MPI_COMM_WORLD);
for ( i=0 ; i < size ; i++ ) {
for ( j=0 ; j < chunk ; j++ ) {
printf("myid=%d,recv from id=%d, data[%d]=%d\n",rank,i,j,rb[i*chunk+j]);
/*打印接收缓冲区从其它进程接收的数据*/
}
}
free(sb);
free(rb);

MPI_Finalize();
}