stapl::skeletons::skeletons_impl::pointer_jumping< Op, Flows > Struct Template Reference

This class abstracts the semantics of pointer jumping skeleton by exposing only the necessary information in its representation. More...

Public Member Functions
	pointer_jumping (Op const &op)
Op	get_op (void) const
std::size_t	operator() (std::size_t a)

Public Types
using	skeleton_tag_type = tags::pointer_jumping
using	base_type = decltype(skeletons::repeat< Flows >(skeletons::elem(skeletons::pointer_jumping_pd(std::declval< Op >())), log_ceil_lazysize< 2 >()))
using	op_type = Op

Detailed Description

template<typename Op, typename Flows>
struct stapl::skeletons::skeletons_impl::pointer_jumping< Op, Flows >

This class abstracts the semantics of pointer jumping skeleton by exposing only the necessary information in its representation.

Pointer jumping skeleton (aka recursive doubling) is used in various computations such as Hillis-Steele scan.

A pointer jumping skeleton consists of log2 levels ( ) of reading from and combining the value with the value already existing at index .

This abstraction not only makes the reconstruction of an pointer_jumping skeleton easier, but also provides access to the underlying operation. Furthermore, it reduces the symbol size for an allreduce skeleton, hence, reducing the total compilation time.

Template Parameters

Op	the operation to be used in each level
Flows	the flows to be used for the repetition

---

The documentation for this struct was generated from the following file:

• pointer_jumping.hpp