5.6.2  Ending Parallel Regions

SummaryAfter finishing the execution of an OpenMP parallel region, the implementation executes ompd_bp_parallel_end.

Format

 
void ompd_bp_parallel_end(void);
  

DescriptionThe OpenMP implementation must execute ompd_bp_parallel_end at every parallel-end event. At the point that the implementation reaches ompd_bp_parallel_end, the binding for ompd_get_curr_parallel_handle is the parallel region that is ending and the binding for ompd_get_curr_task_handle is the task that encountered the parallel construct. After execution of ompd_bp_parallel_end, any parallel_handle that was acquired for the parallel region is invalid and should be released.

Cross References

• parallel construct, see Section 2.6 on page 214.
• ompd_get_curr_parallel_handle, see Section 5.5.6.1 on page 1576.
• ompd_rel_parallel_handle, see Section 5.5.6.4 on page 1583.
• ompd_get_curr_task_handle, see Section 5.5.7.1 on page 1587.