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

#include "omp.h"

#define NELEMENTS 1024*1024*4

int global_NumThreads;

float vecA[NELEMENTS];
float vecB[NELEMENTS];
float vecC[NELEMENTS];

float dp = 0.0;

double time;

int main ()
{

  #pragma omp parallel for z
  for(int i=0; i < NELEMENTS; i ++) {
    vecA[i] = 1.0 ;
    vecB[i] = 2.0 ;
  }

  #pragma omp parallel for schedule(static)
  for(int i=0; i < NELEMENTS; i ++) {
    vecC[i] = vecA[i] * vecB[i];
  }

  
  dp = 0.0;
  time = omp_get_wtime();
  #pragma omp parallel for shared(dp)
  for(int i=0; i < NELEMENTS; i ++) {
    // ustvarimo kritično sekcijo
    #pragma omp critical
    {   // "zakleni ključavnico"
      dp += vecC[i];
    }   // "odkleni ključavnico"
  }
  printf("TIME CRITICAL:   %f        ", omp_get_wtime() - time);
  printf("Dot product = %f \n", dp);
  
  dp = 0.0;
  time = omp_get_wtime();
  #pragma omp parallel for shared(dp)
  for(int i=0; i < NELEMENTS; i ++) {
    // ustvarimo kritično sekcijo
    #pragma omp atomic 
      dp += vecC[i];
  }
  printf("TIME ATOMIC:     %f          ", omp_get_wtime() - time);
  printf("Dot product = %f \n", dp);
  
  dp = 0.0;
  time = omp_get_wtime();
  #pragma omp parallel for reduction(+:dp)
  for(int i=0; i < NELEMENTS; i ++) {
    dp += vecC[i];
  }
  printf("TIME REDUCTION:  %f       ", omp_get_wtime() - time);
  printf("Dot product = %f \n", dp);

  dp = 0.0;
  time = omp_get_wtime();
  for(int i=0; i < NELEMENTS; i ++) {
    dp += vecC[i];
  }
  printf("TIME NOPARALL:  %f       ", omp_get_wtime() - time);
  printf("Dot product = %f \n", dp);
  
  
  
  return 0;
}