Question

I want to add all the iterations that the loop went using the openmp reduction operator.

This is my code:

#include <omp.h>
#include <stdio.h>
#include <stdlib.h>
#define CHUNKSIZE   2
#define N       100

int main (int argc, char *argv[]) 
{
int nthreads, tid, i, chunk;
float a[N], b[N], c[N];

/* Some initializations */
for (i=0; i < N; i++)
  a[i] = b[i] = i * 1.0;
chunk = CHUNKSIZE;


int x=0;

#pragma omp parallel shared(a,b,c,nthreads,chunk) private(i,tid) reduction(+ : x)
  {
  tid = omp_get_thread_num();
  if (tid == 0)
    {
    nthreads = omp_get_num_threads();
    printf("Number of threads = %d\n", nthreads);
    }
  printf("Thread %d starting...\n",tid);

  #pragma omp for schedule(static,chunk)
  for (i=0; i<N; i++)
    {
    c[i] = a[i] + b[i];
    printf("Thread %d: c[%d]= %f\n",tid,i,c[i]);
    x++;
    }

  }  /* end of parallel section */


printf("Value of x is %d" , x);
}

The problem is the final value of x is 100, not 200. I cannot understand the reason why I am not getting the expected value of 200. Could someone please help me?

Answer 1

100 is the expected outcome:

#pragma omp for schedule(static,chunk)
for (i=0; i<N; i++) {
  ...
}

Each thread will get a set of two-element chunks from the [0..N[ interval, and only increment x for the values it got assigned. So the total number of times x++ will be executed, across all threads, is N.

i.e. assuming three threads one will run the loop body for i=0,1,6,7,12,13,..., the second thread for i=2,3,8,9,... and the third for i=4,5,10,11,....