源文件

// mpi_AxB.c

#include "mpi.h"
#include <stdio.h>
#include <stdlib.h>
#include <sys/types.h>
#include <unistd.h>

const int rows = 10;  //the rows of matrix
const int cols = 7;  //the cols of matrix

int main(int argc, char* argv[]){
    int i, j, k, myid, numprocs, anstag;
    int A[rows][cols], B[cols], C[rows];
    int buf[cols], ans,cnt;
    double starttime,endtime;
    double tmp,totaltime;

    MPI_Status status;
    MPI_Init(&argc, &argv);
    MPI_Comm_rank(MPI_COMM_WORLD, &myid);
    MPI_Comm_size(MPI_COMM_WORLD, &numprocs);     

    if(numprocs < 2){
        printf("Error:Too few processes!\n");
        MPI_Abort(MPI_COMM_WORLD,99);
    }

    /* root processor*/
    if(myid == 0){
        /*  initialize A and B  */
        starttime = MPI_Wtime();
        printf("\nA = \n");
        srand(getpid());
        for (i = 0; i < rows; i++){ 
            printf("\t");
            for (j = 0; j < cols; j++)
            {
                A[i][j] = rand()%10;
                printf("%d ",A[i][j]);
            }
            printf("\n");
        }
        printf("\nB = \n");
        for (i = 0; i < cols; i++){
            B[i] = rand()%10;
            printf("\t%d \n",B[i]);
        }
        printf("\n");

        // bcast the B vector to all slave processor
        MPI_Bcast(B, cols, MPI_INT, 0, MPI_COMM_WORLD);

        // partition the A matrix to all slave processor
        for (i = 1; i < numprocs; i++) {
            for (k = i - 1; k < rows; k += numprocs - 1) {
                for (j = 0; j < cols; j++) {    
                    buf[j] = A[k][j];
                }
                MPI_Send(buf, cols, MPI_INT, i, k, MPI_COMM_WORLD);
            }
        }
    }
    /*  slave processor */
    else{
        MPI_Bcast(B, cols, MPI_INT, 0, MPI_COMM_WORLD);

        /* every processor receive the part of A matrix,and make Mul operator with B vector */
        for ( i = myid - 1; i < rows; i += numprocs - 1) {
            MPI_Recv(buf, cols, MPI_INT, 0, i, MPI_COMM_WORLD, &status);
            ans = 0;
    
            for ( j = 0; j < cols; j++) {
                ans += buf[j] * B[j];
            }
            //send back the result
            MPI_Send(&ans, 1, MPI_INT, 0, i, MPI_COMM_WORLD);
        }
    } 

    /* root processor*/
    if(myid == 0){    
        //receive the result from all slave processor
        for ( i = 0; i < rows; i++){
            MPI_Recv(&ans, 1, MPI_INT, MPI_ANY_SOURCE, MPI_ANY_TAG, MPI_COMM_WORLD, &status);
        //sender = status.MPI_SOURCE;
            anstag = status.MPI_TAG;
            C[anstag] = ans;
        }
    }
    endtime = MPI_Wtime();
    totaltime = endtime-starttime;

    /* root processor print C  and timer */
    if(myid == 0){
        printf("\nC = \n");
        for (j = 0; j < rows; j++) {
            printf("\t%d \n",C[j]);
        }
        printf("\n");
        printf("total time:%f s.\n",totaltime);
    }
    MPI_Finalize();
    return 0;
}

编译 链接 运行
用mpicc编译链接

    /* 编译和链接一起 */
    mpicc mpi_AxB.c -o mpi_AxB   //直接生成可执行文件
    
    /* 也可以 编译和链接分开 */
    mpicc -c mpi_AxB.c          //生成.o文件
    mpicc -o mpi_AxB mpi_AxB.o  //进一步生成可执行文件
    
    /* 运行 */
    mpirun ./mpi_AxB

也可以先用g++编译（编译时加上mpi的包含路径），再用mpicc链接。

MPI_INCLUDE = -I/work/soft/mpi/intelmpi/2017.4.239/intel64/include
g++ $(MPI_INCLUDE) -c mpi_AxB.c
mpicc -o mpi_AxB mpi_AxB.o
mpirun ./mpi_AxB

用g++链接会报错

MPI_INCLUDE = -I/work/soft/mpi/intelmpi/2017.4.239/intel64/include
MPILIB = -L/work/soft/mpi/intelmpi/2017.4.239/intel64/bin/lib -lstdc++
g++ $(MPI_INCLUDE) -c mpi_AxB.c
g++ $(MPILIB) -o mpi_AxB mpi_AxB.o
mpirun ./mpi_AxB

运行结果：

A = 
	6 9 9 3 9 7 3 
	6 9 3 1 9 1 3 
	0 0 7 0 1 7 6 
	4 0 4 5 0 2 6 
	5 0 9 3 1 0 6 
	0 7 9 6 8 5 0 
	7 6 5 8 8 3 8 
	9 0 4 3 0 0 1 
	2 2 7 9 4 8 3 
	8 8 9 8 7 1 7 

B = 
	5 
	8 
	7 
	3 
	4 
	2 
	3 


C = 
	233 
	173 
	85 
	85 
	119 
	179 
	204 
	85 
	143 
	242 

Makefile文件：

MPI_INCLUDE = /work/soft/mpi/intelmpi/2017.4.239/intel64/include
MPILIB = -L/work/soft/mpi/intelmpi/2017.4.239/intel64/bin/lib -lstdc++
MPIFLAGS = -I$(MPI_INCLUDE)

有时mpirun需要加一些参数：

ibv_exp_query_device: invalid comp_mask !!! (comp_mask = 0x7f81fdd30040 valid_mask = 0x1)
[r4-node9][[44333,1],23][btl_openib_component.c:1670:init_one_device] error obtaining device attributes for mlx5_0 errno says Invalid argument
--------------------------------------------------------------------------
WARNING: There was an error initializing an OpenFabrics device.

  Local host:   r4-node9
  Local device: mlx5_0
--------------------------------------------------------------------------

报上面这个错，mpirun加参数 -mca pml ucx