Passing submatrices from master to slaves MPI - c

Question

I'm trying to learn MPI and I've run into the following problem in one of my courses:
Consider a matrix A of dimensions n * n in which each element is an integer. Given 2 pair of indices (i1,j1) and (i2,j2) find the submatrix of such dimensions in matrix A for which it's elements sum is maximum.
I'd like some help on how to pass the submatrices to the processes. Should I calculate first how many submatrices (s) are in the matrix and send to each process N/s? How would I send the submatrices?
Some skeleton code I wrote:
#include<mpi.h>
#include<stdio.h>
#include<math.h>
#include<assert.h>
#include<iostream>
using namespace std;
#pragma comment (lib, "msmpi.lib")
enum CommunicationTag
{
COMM_TAG_MASTER_SEND_TASK,
COMM_TAG_MASTER_SEND_TERMINATE,
COMM_TAG_SLAVE_SEND_RESULT,
};
void print_matrix(int mat[10][10], int n) {
for (int i = 0; i < n; i++) {
for (int j = 0; j < n; j++) {
printf("%d ", mat[i][j]);
}
printf("\n");
}
}
int main(int argc, char *argv[]) {
//0. Init part, finding rank and number of processes
int numprocs, rank, rc;
rc = MPI_Init(&argc, &argv);
if (rc != MPI_SUCCESS) {
printf("Error starting MPI program. Terminating \n");
MPI_Abort(MPI_COMM_WORLD, rc);
}
MPI_Comm_size(MPI_COMM_WORLD, &numprocs);
MPI_Comm_rank(MPI_COMM_WORLD, &rank);
printf("I'm rank %d. Num procs %d\n", rank, numprocs); fflush(stdout);
//1. different machine code
if (rank == 0)
{
int n;
scanf("%d", &n);
int i1, i2, j1, j2;
scanf("%d%d%d%d", &i1, &i2, &j1, &j2);
int mat[10][10];
//init data
for (int i = 0; i < n; i++)
for (int j = 0; j < n; j++) {
mat[i][j] = (rand() % 100) - 50; //init random between -50 and 49
}
print_matrix(mat, n);
//here; how do I pass the submatrices to the processes?
for (int i = 1; i < numprocs; i++) {
MPI_Send(&i1, 1, MPI_INT, i, COMM_TAG_MASTER_SEND_TASK, MPI_COMM_WORLD);
MPI_Send(&i2, 1, MPI_INT, i, COMM_TAG_MASTER_SEND_TASK, MPI_COMM_WORLD);
MPI_Send(&j1, 1, MPI_INT, i, COMM_TAG_MASTER_SEND_TASK, MPI_COMM_WORLD);
MPI_Send(&j2, 1, MPI_INT, i, COMM_TAG_MASTER_SEND_TASK, MPI_COMM_WORLD);
//here; how do I pass the submatrices to the processes?
}
}
else {
//if slave ...
}
system("Pause");
}

Answer 1

The first step is to stop thinking about how to use MPI_Send(). The basic solution is to use MPI_Bcast() to transmit A to all the MPI processes.
Then divide the work up (no need to communicate for this, the same dividing logic can run in each process). Compute the sums within each MPI process, and collect them in the main process using MPI_Gather(). Choose the largest and you're done.
It really only requires two MPI operations: Bcast to distribute the input data to all processes, and Gather to centralize the results.
Note that all MPI processes need to execute the collective operations together in lockstep. You only need if (rank == 0) to know which process should load the matrix and analyze the Gathered results.