f-new-features-and-clarifications-in-version-2-0.md

description	Learn more about: F. New features and clarifications in version 2.0
title	F. New features and clarifications in version 2.0
ms.date	01/22/2019
ms.assetid	0d4beb66-f2d5-468c-8cd3-4b00dcbab061

F. New features and clarifications in version 2.0

This appendix summarizes the key changes made to the OpenMP C/C++ specification in moving from version 1.0 to version 2.0. The following items are new features added to the specification:

• Commas are allowed in OpenMP directives.

• Addition of the num_threads clause. This clause allows a user to request a specific number of threads for a parallel construct.

• The threadprivate directive has been extended to accept static block-scope variables.

• C99 Variable Length Arrays are complete types and can be specified anywhere complete types are allowed, such as in the lists of private, firstprivate, and lastprivate clauses (see section 2.7.2).

• A private variable in a parallel region can be marked private again in a nested directive.

• The copyprivate clause has been added. It provides a mechanism to use a private variable to broadcast a value from one member of a team to the other members. It's an alternative to using a shared variable for the value when providing such a shared variable would be difficult (for example, in a recursion requiring a different variable at each level). The copyprivate clause can only appear on the single directive.

• Addition of timing routines omp_get_wtick and omp_get_wtime similar to the MPI routines. These functions are necessary to do wall clock timings.

• An appendix with a list of implementation-defined behaviors in OpenMP C/C++ has been added. An implementation is required to define and document its behavior in these cases.

• The following changes serve to clarify or correct features in the previous OpenMP API specification for C/C++:

  • Clarified that the behavior of omp_set_nested and omp_set_dynamic when omp_in_parallel returns nonzero is undefined.

  • Clarified directive nesting when nested parallel is used.

  • The lock initialization and lock destruction functions can be called in a parallel region.

  • New examples have been added to appendix A.