使用MPI中的MPI_Reduce和MPI_Scatter和mpi

Question

我是mpi中的新手，从这段代码我想将一行2D阵列分配给所有处理器，例如

p1 = 1,2,3,4
p2 = 5,6,7,8
p3 = 9,10,11,12
p4 = 13,14,15,16

使用(mpirun -np 4 ./a)运行程序时，     MPI_Scatter 工作正常但是     MPI_Reduce 导致终端停止我不知道如何通过找到localmax或（mymax）来处理MPI_Reduce。任何人都可以帮忙吗？

#include "mpi.h"
#include <stdio.h>
#define size 4

int main ()
{
    int np, rank, sendcount, recvcount, source,i;
    int recvbuf[size];
    int mymax;
    int max=0;

    MPI_Init(NULL,NULL);
    MPI_Comm_rank(MPI_COMM_WORLD, &rank);
    MPI_Comm_size(MPI_COMM_WORLD, &np);

    int sendbuf[size][size] ={
    {1, 2, 3, 4},
    {5, 6, 7, 8},
    {9,10,11, 12},
    {13, 14, 15, 16}};
    source = 1;
    sendcount = size;
    recvcount = size;

    MPI_Scatter(sendbuf,sendcount,MPI_INT,recvbuf,recvcount,MPI_INT,source,MPI_COMM_WORLD);
    printf("rank= %d  Results: %d %d %d %d \n",rank,recvbuf[0],recvbuf[1],recvbuf[2],recvbuf[3]);

    //Each processor has a row, now find local max
    mymax = recvbuf[0]; 
    for(i=0;i<recvcount;i++) {
        if(mymax < recvbuf[i]) {
            mymax = recvbuf[i];
        }
    }

    MPI_Reduce(&mymax,&max ,1,MPI_INT, MPI_MAX, 0, MPI_COMM_WORLD);

    if (rank==0) {
        printf(" Processor %d has max data after reduce : max= %d ", rank,max);
    }
    else 
        printf("----.\n");

    MPI_Finalize();
}