PWD009

PWD009: Incorrect privatization in OpenMP parallel region

Issue

A variable is being incorrectly privatized in the OpenMP datascoping clauses.

Actions

Change the data scope of the variable from private to shared.

Relevance

Specifying an invalid scope for a variable may introduce race conditions and produce incorrect results. For instance, when a variable must be shared among threads but it is privatized instead.

Code example

In the following code, C is incorrectly privatized:

void example(int m, double *A, double *B, double *C) {
  double temp;

  #pragma omp parallel for default(none) private(temp, C) shared(A, B, m)
  for (int i = 0; i < m; i++) {
    temp = A[i] * B[i];
    C[i] = C[i] + temp;
  }
}

To fix this, it should be moved to the shared clause:

void example(int m, double *A, double *B, double *C) {
  double temp;

  #pragma omp parallel for default(none) private(temp) shared(A, B, C, m)
  for (int i = 0; i < m; i++) {
    temp = A[i] * B[i];
    C[i] = C[i] + temp;
  }
}

Related resources

• PWD009 examples

References

• Data-Sharing Attribute Clauses - OPENMP API Specification: Version 5.0 November 2018 [last checked August 2021]