matrix multiplication using Mpi_Scatter and Mpi_Gather

Question

I newbie to mpi programming. I was trying to write matrix multiplication. Went through the post MPI Matrix Multiplication with scatter gather about matrix multiplication usi

Answer 1

Call to mpi_finalize doesn't indicate that all the MPI processes are terminated like in OpenMP !

In most of mpi implementation, all the processes execute the instruction before the MPI_init and after MPI_Finalized.

A good practice is to do nothing before MPI_Init and after MPI_Finalized.

Answer 2

The result matrix c is getting printed by all processes because every process executes the function void print_results(char *prompt, int a[N][N]). Since you are gathering at the process having rank 0, add a statement if (rank == 0) before calling the print_results(...) function. Further, the result is incorrect because of a wrong loop logic in :

                for (j = 0; j < N; j++)
                {
                        sum = sum + aa[j] * b[i][j];                
                }

This should be :

                for (j = 0; j < N; j++)
                {
                        sum = sum + aa[j] * b[j][i];                
                }

Also there is no need to broadcast b as all processes already already have a copy of it and you can avoid MPI_Barrier(). The complete program then becomes :

#define N 4
#include <stdio.h>
#include <math.h>
#include <sys/time.h>
#include <stdlib.h>
#include <stddef.h>
#include "mpi.h"


void print_results(char *prompt, int a[N][N]);

int main(int argc, char *argv[])
{
    int i, j, k, rank, size, tag = 99, blksz, sum = 0;
    int a[N][N]={{1,2,3,4},{5,6,7,8},{9,1,2,3},{4,5,6,7,}};
    int b[N][N]={{1,2,3,4},{5,6,7,8},{9,1,2,3},{4,5,6,7,}};
    int c[N][N];
    int aa[N],cc[N];

    MPI_Init(&argc, &argv);
    MPI_Comm_size(MPI_COMM_WORLD, &size);
    MPI_Comm_rank(MPI_COMM_WORLD, &rank);

    //scatter rows of first matrix to different processes     
    MPI_Scatter(a, N*N/size, MPI_INT, aa, N*N/size, MPI_INT,0,MPI_COMM_WORLD);

    //broadcast second matrix to all processes
    MPI_Bcast(b, N*N, MPI_INT, 0, MPI_COMM_WORLD);

    MPI_Barrier(MPI_COMM_WORLD);

          //perform vector multiplication by all processes
          for (i = 0; i < N; i++)
            {
                    for (j = 0; j < N; j++)
                    {
                            sum = sum + aa[j] * b[j][i];  //MISTAKE_WAS_HERE               
                    }
                    cc[i] = sum;
                    sum = 0;
            }

    MPI_Gather(cc, N*N/size, MPI_INT, c, N*N/size, MPI_INT, 0, MPI_COMM_WORLD);

    MPI_Barrier(MPI_COMM_WORLD);        
    MPI_Finalize();
    if (rank == 0)                         //I_ADDED_THIS
        print_results("C = ", c);
}

void print_results(char *prompt, int a[N][N])
{
    int i, j;

    printf ("\n\n%s\n", prompt);
    for (i = 0; i < N; i++) {
            for (j = 0; j < N; j++) {
                    printf(" %d", a[i][j]);
            }
            printf ("\n");
    }
    printf ("\n\n");
}

Then c =

C = 
 54 37 47 57

 130 93 119 145

 44 41 56 71

 111 79 101 123