MPI Matrix Multiplication Times 3 Due Write 3 MPI Programs

MPI Matrix Multiplication Times 3 - Due Write 3 MPI Programs for Matrix Multiplication Assume all process start with initial portion of data and hold results in the end 2 Version One: All to all broudcast Version Two: Row/Column broudcast Version Three: Cannon\'s Alogorithm Compare Run on 16 processes With Matrix size N=256 Compare running time of each Deliverables Source Code Analysis of running time Graph of time

Solution

#include \"stdio.h\"

#include \"mpi.h\"

#define N    500        /* number of rows and columns in matrix */

MPI_Status status;

double a[N][N],b[N][N],c[N][N]; //matrix used

      

main(int argc, char **argv){

struct timeval start, stop;

int numberOfTasks,

mtype,

taskID,

numberOfWorkers,

source,

destination,

rows,

averageRow,

extra,

offset,i,j,k;

//first initialization

MPI_Init(&argc, &argv);

MPI_Comm_rank(MPI_COMM_WORLD, &taskID);

MPI_Comm_size(MPI_COMM_WORLD, &numberOfTasks);

numberOfWorkers = numberOfTasks-1;

//---------------------------- master ----------------------------//

if (taskID == 0) {

   for (i=0; i<N; i++) {

      for (j=0; j<N; j++) {  

a[i][j]= 1.0;

b[i][j]= 2.0;

     }

          }

    /* send matrix data to the worker tasks */

   gettimeofday(&start, 0);

          averageRow = N/numberOfWorkers; //average rows per worker

   extra= N%numberOfWorkers; //extra rows

          offset = 0;

   

          for (destination=1; destination<=numberOfWorkers; destination++){      

     if(destination<=extra){

        rows = averageRow+1;

     }else{

        rows = averageRow;

     }

     mtype = 1;

     MPI_Send(&offset, 1, MPI_INT, destination, mtype, MPI_COMM_WORLD);

     MPI_Send(&rows, 1, MPI_INT, destination, mtype, MPI_COMM_WORLD);

     MPI_Send(&a[offset][0], rows*N, MPI_DOUBLE,destination,mtype, MPI_COMM_WORLD);

     MPI_Send(&b, N*N, MPI_DOUBLE, destination, mtype, MPI_COMM_WORLD);

     offset = offset + rows;

}

    /* wait for results from all worker tasks */

for (i=1; i<=numberOfWorkers; i++){

     mtype = 2;

     source = i;

     MPI_Recv(&offset, 1, MPI_INT, source, mtype, MPI_COMM_WORLD, &status);

     MPI_Recv(&rows, 1, MPI_INT, source, mtype, MPI_COMM_WORLD, &status);

     MPI_Recv(&c[offset][0], rows*N, MPI_DOUBLE, source, mtype, MPI_COMM_WORLD, &status);

}

gettimeofday(&stop, 0);

printf(\"Upper Left = %6.2f    Upper Right = %6.2f \ \",c[0][0],c[0][N-1]);

printf(\"Lower Left = %6.2f    Lower Right = %6.2f \ \",c[N-1][0],c[N-1][N-1]);

fprintf(stdout,\"Time = %.6f\ \ \",(stop.tv_sec+stop.tv_usec*1e-6)-(start.tv_sec+start.tv_usec*1e-6));

}

/*---------------------------- worker----------------------------*/

if (taskID > 0) {

     source = 0;

     mtype = 1;

     MPI_Recv(&offset, 1, MPI_INT, source, mtype, MPI_COMM_WORLD, &status);

     MPI_Recv(&rows, 1, MPI_INT, source, mtype, MPI_COMM_WORLD, &status);

     MPI_Recv(&a, rows*N, MPI_DOUBLE, source, mtype, MPI_COMM_WORLD, &status);

     MPI_Recv(&b, N*N, MPI_DOUBLE, source, mtype, MPI_COMM_WORLD, &status);

    /* Matrix multiplication */

     for (k=0; k<N; k++)

   for (i=0; i<rows; i++) {

            c[i][k] = 0.0;

     for (j=0; j<N; j++)

   c[i][k] = c[i][k] + a[i][j] * b[j][k];

         }

     mtype = 2;

     MPI_Send(&offset, 1, MPI_INT, 0, mtype, MPI_COMM_WORLD);

     MPI_Send(&rows, 1, MPI_INT, 0, mtype, MPI_COMM_WORLD);

     MPI_Send(&c, rows*N, MPI_DOUBLE, 0, mtype, MPI_COMM_WORLD);

}

   

MPI_Finalize();

}