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

#define TIMES 40000000

float randx(void)
{
   float v;
   srand((unsigned int)time(NULL));
   v =  (float)rand()/(float)(RAND_MAX);
   return v;
}

int main(void) {

	int comm_sz, my_rank;
	int i, local_n, local_t, global_t = 0;

	MPI_Init(NULL, NULL);

	/* 
	   Since there is a constant defining how many times the cycle has to be executed in total, 
	   it would not be necessary to have the rank 0 process bradcasting the number of iterations
	   to be executed by each process.
	*/
	MPI_Comm_size(MPI_COMM_WORLD, &comm_sz);
	MPI_Comm_rank(MPI_COMM_WORLD, &my_rank);
	local_n = TIMES / comm_sz; 	// static distribution of work, 
	                           // knowing that the remain of the division is zero.
	
	local_t = 0; 
	for( i = 0; i < local_n; i++ ){
		double x, y;
 		x = randx( );  /*randx return a random real number 
                  		between 0 and 1*/ 
   		y = randx( );	
   		if( x*x+y*y<=1.0) local_t = local_t + 1;
	} 
  	/* char processor_name[MPI_MAX_PROCESSOR_NAME];
  	int name_len;
  	MPI_Get_processor_name(processor_name, &name_len);
  	printf("Processor %s, rank %d"
         " out of %d processors, local_t:%d\n",
		 processor_name, my_rank, comm_sz, local_t); 
	*/

	MPI_Reduce(&local_t, &global_t, 1, MPI_INT, MPI_SUM, 0, MPI_COMM_WORLD);
	

	if(my_rank == 0) {
		double pi;
		pi = 4.0*((double)global_t/(double)TIMES);
		//printf("global_t:%d pi:%f\n", global_t, pi);
		printf("pi:%f\n", pi);
	}

	MPI_Finalize();

	return 0;
}