Intel® oneAPI DPC++/C++ Compiler Developer Guide and Reference

ID 767253
Date 7/13/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

OpenMP* Examples

The following examples show how to use OpenMP* features.

A Simple Difference Operator

This example shows a simple parallel loop where the amount of work in each iteration is different. Dynamic scheduling is used to improve load balancing.

The for pragma has a nowait clause because there is an implicit barrier at the end of the parallel region. Therefore it is not necessary to also have a barrier at the end of the for region.

void for1(float a[], float b[], int n) {
  int i, j;
  #pragma omp parallel shared(a,b,n) 
  {
   #pragma omp for schedule(dynamic,1) private (i,j) nowait
    for (i = 1; i < n; i++)
       for (j = 0; j < i; j++)
         b[j + n*i] = (a[j + n*i] + a[j + n*(i-1)]) / 2.0;
  } 
}

Two Difference Operators: for Loop Version

This example uses two parallel loops fused to reduce fork/join overhead. The first for pragma has a nowait clause because all the data used in the second loop is different than all the data used in the first loop.

void for2(float a[], float b[], float c[], float d[], int n, int m) {
  int i, j;
  #pragma omp parallel shared(a,b,c,d,n,m) private(i,j) 
  {
    #pragma omp for schedule(dynamic,1) nowait
    for (i = 1; i < n; i++)
      for (j = 0; j < i; j++)
        b[j + n*i] = ( a[j + n*i] + a[j + n*(i-1)] )/2.0;
    #pragma omp for schedule(dynamic,1) nowait
    for (i = 1; i < m; i++)
      for (j = 0; j < i; j++)
        d[j + m*i] = ( c[j + m*i] + c[j + m*(i-1)] )/2.0;
  } 
}

Two Difference Operators: sections Version

This example demonstrates the use of the sections pragma . The logic is identical to the preceding for pragma example, but uses a sections pragma instead of a for pragma . Here the speedup is limited to two because there are only two units of work whereas in the example above there are (n-1) + (m-1) units of work.

void sections1(float a[], float b[], float c[], float d[], int n, int m) {
  int i, j;
  #pragma omp parallel shared(a,b,c,d,n,m) private(i,j) 
  {
    #pragma omp sections nowait 
    {
      #pragma omp section
       for (i = 1; i < n; i++)
         for (j = 0; j < i; j++)
           b[j + n*i] = ( a[j + n*i] + a[j + n*(i-1)] )/2.0;
      #pragma omp section
       for (i = 1; i < m; i++)
         for (j = 0; j < i; j++)
           d[j + m*i] = ( c[j + m*i] + c[j + m*(i-1)] )/2.0;
     }
   } 
}

Update a Shared Scalar

This example demonstrates how to use a single construct to update an element of the shared array a. The optional nowait clause after the first loop is omitted because it is necessary to wait at the end of the loop before proceeding into the single construct to avoid a race condition.

void sp_1a(float a[], float b[], int n) {
  int i;
  #pragma omp parallel shared(a,b,n) private(i) 
  {
    #pragma omp for
      for (i = 0; i < n; i++)
        a[i] = 1.0 / a[i];
      #pragma omp single
        a[0] = MIN( a[0], 1.0 );
      #pragma omp for nowait
      for (i = 0; i < n; i++)
      b[i] = b[i] / a[i];
   } 
}

More Samples

Additional OpenMP code samples for the Intel® oneAPI DPC++/C++ Compiler are available in the oneAPI Samples GitHub* repository.