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

#define N 3

int main(int argc, char *argv[]){

	int size, rank, sum = 0, result = 0, i, j, cij, recv_rank;
	int row[N], col[N];

	int a [N][N] = {1,2,3,
					4,5,6,
					7,8,9};

	int b [N][N] = {1,2,3,
					4,5,6,
					7,8,9};

	int c [N][N]; //Result {30,36,42,66,81,96,102,126,150}

	MPI_Datatype mytype;

	MPI_Status status;
	MPI_Init(NULL, NULL);
	MPI_Comm_size(MPI_COMM_WORLD, &size);
	MPI_Comm_rank(MPI_COMM_WORLD, &rank);

	if(size != (N*N)){
		printf("Número de processos diferente da ordem da matriz!!!\n");
		exit(0);
	}

	if(rank == 0){

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

		printf("\nMatriz B\n");
		for(i = 0; i < N; i++){
			for(j = 0; j < N; j++){
				printf("[%d]\t", b[i][j]);
			}
			printf("\n");
		}

		int _ij;


		MPI_Type_vector(N, 1, N, MPI_INT, &mytype);
		MPI_Type_commit(&mytype);

		for(i = 0; i < N; i++){
			for(j = 0; j < N; j++){
				_ij = i*N+j;
				if((_ij)!=0){
					MPI_Send(&a[i][0], N, MPI_INT, _ij, 0, MPI_COMM_WORLD);
					MPI_Send(&b[0][j], 1, mytype, _ij, 0, MPI_COMM_WORLD);
				}
			}
			col[i] = b[0][i];
			row[i] = a[i][0];
		}

		/*for(i = 0; i < N; i++){
			printf("r = [%d]\n", row[i]);
			printf("c = [%d]\n", col[i]);
		}*/

		//Recv 1 to n-1 processos
		for (i = 1; i < size; i++){
			MPI_Recv(&cij, 1, MPI_INT, i, 0, MPI_COMM_WORLD, &status);
			c[(i/N)][(i%N)] = cij;
		}

		#pragma omp parallel reduction (+: sum) num_threads(N)
		{
			int thread_number = omp_get_thread_num();
			sum = (row[thread_number] * col[thread_number]);
		}
		c[0][0] = sum;

		printf("\nResult: \n");

		for(i = 0; i < N; i++){
			for(j = 0; j < N; j++){
				printf("[%d]\t", c[i][j]);
			}
			printf("\n");
		}
		printf("\n");
	} else {
		// 1 até n-1 processos
		//printf("My Rank %d\n",rank);

		MPI_Recv(&row, N, MPI_INT, 0, 0, MPI_COMM_WORLD, &status);
		MPI_Recv(&col, N, MPI_INT, 0, 0, MPI_COMM_WORLD, &status);

		/*for(i = 0; i < N; i++){
			printf("My Rank %d r = [%d]\n", rank, row[i]);
		}

		for(i = 0; i < N; i++){
			printf("My Rank %d c = [%d]\n", rank, col[i]);
		}*/

		#pragma omp parallel reduction (+: sum) num_threads(N)
		{
			int thread_number = omp_get_thread_num();
			sum = (row[thread_number] * col[thread_number]);
		}
		//printf("From Rank %d c[%d][%d] = %d\n", rank, (rank/N),(rank%N), sum);
		MPI_Send(&sum, 1, MPI_INT, 0, 0, MPI_COMM_WORLD);
	}
	MPI_Finalize();
	return 0;
}