LV5Z1

// https://github.com/elahehrashedi/MPI_Matrix_Multiplication_Scatter_Gather/blob/master/matrix.c

#include <mpi.h>
#include <stdio.h>
#include <stdlib.h>

int main(int argc, char *argv[])
{
    int matrixSize, rank, size;

    double *sendA, *sendB, *finalC, *recA, *recC;

    MPI::Init(argc, argv);

    size = MPI::COMM_WORLD.Get_size();
    rank = MPI::COMM_WORLD.Get_rank();

    if (rank == 0)
    {
        matrixSize = 10;
        printf("rows is %d \n", matrixSize);

        sendA = new double[matrixSize * matrixSize];
        sendB = new double[matrixSize * matrixSize];
        finalC = new double[matrixSize * matrixSize];

        for (int i = 0; i < matrixSize; i++)
            for (int j = 0; j < matrixSize; j++)
            {
                sendA[i * matrixSize + j] = i + 1;
            }
        for (int i = 0; i < matrixSize; i++)
            for (int j = 0; j < matrixSize; j++)
            {
                sendB[i * matrixSize + j] = j + 1;
            }
    }

    MPI::COMM_WORLD.Barrier();
    MPI::COMM_WORLD.Bcast(&matrixSize, 1, MPI::INT, 0);

    recA = new double[matrixSize / size * matrixSize];
    if (sendB == NULL)
    {
        sendB = new double[matrixSize * matrixSize];
    }
    recC = new double[matrixSize / size * matrixSize];

    // now p0 scatter matrix sendA to all
    MPI::COMM_WORLD.Scatter(sendA, matrixSize * matrixSize / size, MPI::DOUBLE, recA, matrixSize * matrixSize / size, MPI::DOUBLE, 0);

    //now p0 broadcast sendB to all others
    MPI::COMM_WORLD.Bcast(sendB, matrixSize * matrixSize, MPI::DOUBLE, 0);

    // all do this part to calculate recC as multiplicated matrix
    for (int i = 0; i < (matrixSize / size); i++)
        for (int j = 0; j < matrixSize; j++)
        {
            recC[i * matrixSize + j] = 0;
            for (int k = 0; k < matrixSize; k++)
                recC[i * matrixSize + j] += recA[i * matrixSize + k] * sendB[k * matrixSize + j];
        }

    MPI::COMM_WORLD.Barrier();
    MPI::COMM_WORLD.Gather(recC, matrixSize * matrixSize / size, MPI::DOUBLE, finalC, matrixSize * matrixSize / size, MPI::DOUBLE, 0);
    MPI::COMM_WORLD.Barrier();
    if (rank == 0)
    {
        for (int i = 0; i < matrixSize; i++)
        {
            for (int j = 0; j < matrixSize; j++)
                printf(" %.2f ", finalC[i * matrixSize + j]);
            printf("\n");
        }
    }

    MPI::Finalize();
    return 0;
}