Execution Policies

The standard defines three execution policies:

• std::execution::sequenced_policy
• std::execution::parallel_policy
• std::execution::parallel_unsequenced_policy

The corresponding policy tag specifies whether a program should run sequentially, in parallel, or parallel with vectorisation.

• std::execution::seq: runs the program sequentially
• std::execution::par: runs the program in parallel on multiple threads
• std::execution::par_unseq: runs the program in parallel on multiple threads and allows the interleaving of individual loops; permits a vectorised version with SIMD (Single Instruction Multiple Data) extensions.

The usage of the execution policy std::execution::par or std::execution::par_unseq allows the algorithm to run parallel or parallel and vectorised. This policy is a permission and not a requirement.

The following code snippet applies all execution policies.

 1 std::vector<int> v = {1, 2, 3, 4, 5, 6, 7, 8, 9...