OpenMP array indexing in nested loop

Question

I'm having an issue trying to parallelize a nested loop using OpenMP - it's just demo code for playing around and getting used to.

int* myresults = new int[1000]   
#pragma omp parallel
{
    #pragma omp for 
    for(int z=0; z<=mainCount;z++)
    {
        results[n++] = myfunc(z,0); //compute something


        for(int i=0;i<=secondCount;i+=5)
        {
            results[n++]=myfunc(z,i);
        }
    }
}

My problem is the indexing of my results array. I assume since OpenMP is parallelizing the first for-loop that he is using positions of results[] twice or more because results[n++] will produce undefined behaviour (as it cannot be assured that n is incremented right), am I correct?

How can I correctly index and store my results?

Hristo Iliev · Accepted Answer · 2012-06-04 06:26:41Z

1

n is incremented once for myfunc(z,0) and 1 + secondCount/5 times for myfunc(z,i). This means that n should be z * (2 + secondCount/5) at the beginning of each iteration of the outer loop. You should rewrite your code like this:

#pragma omp parallel for private(n)
for (int z = 0; z <= mainCount; z++)
{
    n = z * (2 + secondCount/5);

    results[n++] = myfunc(z, 0);

    for (int i = 0; i <= secondCount; i += 5)
        results[n++] = myfunc(z, i);
}

answered Jun 4, 2012 at 6:26

Hristo Iliev

74.8k12 gold badges148 silver badges197 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

OpenMP array indexing in nested loop

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related