zadatak

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

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

    double *sendA, *sendB, *finalC, *recA, *recC;
    //definirnje varijabli za mjerenje userCPU i realCPU
    //double t1, t2;
    MPI::Init(argc, argv);

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

    if (rank == 0)
    {
        matrixSize = 4;
        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);
    //Start time t1
    //t1 = MPI::Wtime();
    recA = new double[matrixSize / size * matrixSize];
    if (sendB == NULL)
    {
        sendB = new double[matrixSize * matrixSize];
    }
    recC = new double[matrixSize / size * matrixSize];


    MPI::COMM_WORLD.Scatter(sendA, matrixSize * matrixSize / size, MPI::DOUBLE, recA, matrixSize * matrixSize / size, MPI::DOUBLE, 0);


    MPI::COMM_WORLD.Bcast(sendB, matrixSize * matrixSize, MPI::DOUBLE, 0);


    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();
    //End time t2
    //t2 = MPI::Wtime();

    /*
        IZRACUN realCPU Time
        double totalTime;
        double elapsedTime = t2-t1;
        MPI::COMM_WORLD.Reduce( &elapsedTime, &totalTime, 1, MPI::DOUBLE, MPI::SUM, 0 );
    */

    if (rank == 0)
    {
        /*
        IZRACUN MLFPOS
        int n_flop = 2 * matrixSize * matrixSize * matrixSize;
        double time = t2- t1;
        double mflops = n_flop / (time * 1000000);
        */
        for (int i = 0; i < matrixSize; i++)
        {
            for (int j = 0; j < matrixSize; j++)
                printf(" %.2f ", finalC[i * matrixSize + j]);
            printf("\n");
        }
        printf("Send A: \n");
        for (int i = 0; i < matrixSize; i++)
        {
            for (int j = 0; j < matrixSize; j++)
                printf(" %.2f ", sendA[i * matrixSize + j]);
            printf("\n");
        }
        printf("Send B: \n");
        for (int i = 0; i < matrixSize; i++)
        {
            for (int j = 0; j < matrixSize; j++)
                printf(" %.2f ", sendB[i * matrixSize + j]);
            printf("\n");
        }
    }

    MPI::Finalize();
    return 0;
}