hpx/parallel/container_algorithms/exclusive_scan.hpp#

See Public API for a list of names and headers that are part of the public HPX API.

namespace hpx

Top-level namespace.

namespace ranges

Functions

template<typename InIter, typename Sent, typename OutIter, typename T = typename std::iterator_traits<InIter>::value_type, typename Op = std::plus<T>>
exclusive_scan_result<InIter, OutIter> exclusive_scan(InIter first, Sent last, OutIter dest, T init, Op &&op = Op())#

Assigns through each iterator i in [result, result + (last - first)) the value of GENERALIZED_NONCOMMUTATIVE_SUM(binary_op, init, *first, …, *(first + (i - result) - 1)).

The difference between exclusive_scan and inclusive_scan is that inclusive_scan includes the ith input element in the ith sum. If op is not mathematically associative, the behavior of inclusive_scan may be non-deterministic.

Note

Complexity: O(last - first) applications of the predicate op.

Note

GENERALIZED_NONCOMMUTATIVE_SUM(op, a1, …, aN) is defined as:

  • a1 when N is 1

  • op(GENERALIZED_NONCOMMUTATIVE_SUM(op, a1, …, aK), GENERALIZED_NONCOMMUTATIVE_SUM(op, aM, …, aN)) where 1 < K+1 = M <= N.

Template Parameters
  • FwdIter1 – The type of the source iterators used (deduced). This iterator type must meet the requirements of an forward iterator.

  • Sent – The type of the source sentinel (deduced). This sentinel type must be a sentinel for FwdIter1.

  • FwdIter2 – The type of the iterator representing the destination range (deduced). This iterator type must meet the requirements of an forward iterator.

  • T – The type of the value to be used as initial (and intermediate) values (deduced).

  • Op – The type of the binary function object used for the reduction operation.

Parameters
  • first – Refers to the beginning of the sequence of elements the algorithm will be applied to.

  • last – Refers to sentinel value denoting the end of the sequence of elements the algorithm will be applied.

  • dest – Refers to the beginning of the destination range.

  • init – The initial value for the generalized sum.

  • op – Specifies the function (or function object) which will be invoked for each of the values of the input sequence. This is a binary predicate. The signature of this predicate should be equivalent to:

    Ret fun(const Type1 &a, const Type1 &b);
    
    The signature does not need to have const&, but the function must not modify the objects passed to it. The types Type1 and Ret must be such that an object of a type as given by the input sequence can be implicitly converted to any of those types.

Returns

The exclusive_scan algorithm returns an input iterator to the point denoted by the sentinel and an output iterator to the element in the destination range, one past the last element copied.

template<typename ExPolicy, typename FwdIter1, typename Sent, typename FwdIter2, typename T = typename std::iterator_traits<FwdIter1>::value_type, typename Op = std::plus<T>>
parallel::util::detail::algorithm_result<ExPolicy, exclusive_scan_result<FwdIter1, FwdIter2>>::type exclusive_scan(ExPolicy &&policy, FwdIter1 first, Sent last, FwdIter2 dest, T init, Op &&op = Op())#

Assigns through each iterator i in [result, result + (last - first)) the value of GENERALIZED_NONCOMMUTATIVE_SUM(binary_op, init, *first, …, *(first + (i - result) - 1)).

The reduce operations in the parallel exclusive_scan algorithm invoked with an execution policy object of type sequenced_policy execute in sequential order in the calling thread.

The reduce operations in the parallel exclusive_scan algorithm invoked with an execution policy object of type parallel_policy or parallel_task_policy are permitted to execute in an unordered fashion in unspecified threads, and indeterminately sequenced within each thread.

The difference between exclusive_scan and inclusive_scan is that inclusive_scan includes the ith input element in the ith sum. If op is not mathematically associative, the behavior of inclusive_scan may be non-deterministic.

Note

Complexity: O(last - first) applications of the predicate op.

Note

GENERALIZED_NONCOMMUTATIVE_SUM(op, a1, …, aN) is defined as:

  • a1 when N is 1

  • op(GENERALIZED_NONCOMMUTATIVE_SUM(op, a1, …, aK), GENERALIZED_NONCOMMUTATIVE_SUM(op, aM, …, aN)) where 1 < K+1 = M <= N.

Template Parameters
  • ExPolicy – The type of the execution policy to use (deduced). It describes the manner in which the execution of the algorithm may be parallelized and the manner in which it executes the assignments.

  • FwdIter1 – The type of the source iterators used (deduced). This iterator type must meet the requirements of an forward iterator.

  • Sent – The type of the source sentinel (deduced). This sentinel type must be a sentinel for FwdIter1.

  • FwdIter2 – The type of the iterator representing the destination range (deduced). This iterator type must meet the requirements of an forward iterator.

  • T – The type of the value to be used as initial (and intermediate) values (deduced).

  • Op – The type of the binary function object used for the reduction operation.

Parameters
  • policy – The execution policy to use for the scheduling of the iterations.

  • first – Refers to the beginning of the sequence of elements the algorithm will be applied to.

  • last – Refers to sentinel value denoting the end of the sequence of elements the algorithm will be applied.

  • dest – Refers to the beginning of the destination range.

  • init – The initial value for the generalized sum.

  • op – Specifies the function (or function object) which will be invoked for each of the values of the input sequence. This is a binary predicate. The signature of this predicate should be equivalent to:

    Ret fun(const Type1 &a, const Type1 &b);
    
    The signature does not need to have const&, but the function must not modify the objects passed to it. The types Type1 and Ret must be such that an object of a type as given by the input sequence can be implicitly converted to any of those types.

Returns

The exclusive_scan algorithm returns a hpx::future<util::in_out_result<FwdIter1, FwdIter2>> if the execution policy is of type sequenced_task_policy or parallel_task_policy and returns util::in_out_result<FwdIter1, FwdIter2> otherwise. The exclusive_scan algorithm returns an input iterator to the point denoted by the sentinel and an output iterator to the element in the destination range, one past the last element copied.

template<typename Rng, typename O, typename T = typename std::iterator_traits<hpx::traits::range_iterator_t<Rng>>::value_type, typename Op = std::plus<T>>
exclusive_scan_result<traits::range_iterator_t<Rng>, O> exclusive_scan(Rng &&rng, O dest, T init, Op &&op = Op())#

Assigns through each iterator i in [result, result + (last - first)) the value of GENERALIZED_NONCOMMUTATIVE_SUM(+, init, *first, …, *(first + (i - result) - 1))

The difference between exclusive_scan and inclusive_scan is that inclusive_scan includes the ith input element in the ith sum.

Note

Complexity: O(last - first) applications of the predicate std::plus<T>.

Note

GENERALIZED_NONCOMMUTATIVE_SUM(+, a1, …, aN) is defined as:

  • a1 when N is 1

  • GENERALIZED_NONCOMMUTATIVE_SUM(+, a1, …, aK)

    • GENERALIZED_NONCOMMUTATIVE_SUM(+, aM, …, aN) where 1 < K+1 = M <= N.

Template Parameters
  • Rng – The type of the source range used (deduced). The iterators extracted from this range type must meet the requirements of an forward iterator.

  • O – The type of the iterator representing the destination range (deduced). This iterator type must meet the requirements of an forward iterator.

  • T – The type of the value to be used as initial (and intermediate) values (deduced).

  • Op – The type of the binary function object used for the reduction operation.

Parameters
  • rng – Refers to the sequence of elements the algorithm will be applied to.

  • dest – Refers to the beginning of the destination range.

  • init – The initial value for the generalized sum.

  • op – Specifies the function (or function object) which will be invoked for each of the values of the input sequence. This is a binary predicate. The signature of this predicate should be equivalent to:

    Ret fun(const Type1 &a, const Type1 &b);
    
    The signature does not need to have const&, but the function must not modify the objects passed to it. The types Type1 and Ret must be such that an object of a type as given by the input sequence can be implicitly converted to any of those types.

Returns

The exclusive_scan algorithm returns an input iterator to the point denoted by the sentinel and an output iterator to the element in the destination range, one past the last element copied.

template<typename ExPolicy, typename Rng, typename O, typename T = typename std::iterator_traits<hpx::traits::range_iterator_t<Rng>>::value_type, typename Op = std::plus<T>>
parallel::util::detail::algorithm_result<ExPolicy, exclusive_scan_result<traits::range_iterator_t<Rng>, O>>::type exclusive_scan(ExPolicy &&policy, Rng &&rng, O dest, T init, Op &&op = Op())#

Assigns through each iterator i in [result, result + (last - first)) the value of GENERALIZED_NONCOMMUTATIVE_SUM(+, init, *first, …, *(first + (i - result) - 1))

The reduce operations in the parallel exclusive_scan algorithm invoked with an execution policy object of type sequenced_policy execute in sequential order in the calling thread.

The reduce operations in the parallel exclusive_scan algorithm invoked with an execution policy object of type parallel_policy or parallel_task_policy are permitted to execute in an unordered fashion in unspecified threads, and indeterminately sequenced within each thread.

The difference between exclusive_scan and inclusive_scan is that inclusive_scan includes the ith input element in the ith sum.

Note

Complexity: O(last - first) applications of the predicate std::plus<T>.

Note

GENERALIZED_NONCOMMUTATIVE_SUM(+, a1, …, aN) is defined as:

  • a1 when N is 1

  • GENERALIZED_NONCOMMUTATIVE_SUM(+, a1, …, aK)

    • GENERALIZED_NONCOMMUTATIVE_SUM(+, aM, …, aN) where 1 < K+1 = M <= N.

Template Parameters
  • ExPolicy – The type of the execution policy to use (deduced). It describes the manner in which the execution of the algorithm may be parallelized and the manner in which it executes the assignments.

  • Rng – The type of the source range used (deduced). The iterators extracted from this range type must meet the requirements of an forward iterator.

  • O – The type of the iterator representing the destination range (deduced). This iterator type must meet the requirements of an forward iterator.

  • T – The type of the value to be used as initial (and intermediate) values (deduced).

  • Op – The type of the binary function object used for the reduction operation.

Parameters
  • policy – The execution policy to use for the scheduling of the iterations.

  • rng – Refers to the sequence of elements the algorithm will be applied to.

  • dest – Refers to the beginning of the destination range.

  • init – The initial value for the generalized sum.

  • op – Specifies the function (or function object) which will be invoked for each of the values of the input sequence. This is a binary predicate. The signature of this predicate should be equivalent to:

    Ret fun(const Type1 &a, const Type1 &b);
    
    The signature does not need to have const&, but the function must not modify the objects passed to it. The types Type1 and Ret must be such that an object of a type as given by the input sequence can be implicitly converted to any of those types.

Returns

The exclusive_scan algorithm returns a hpx::future<util::in_out_result <traits::range_iterator_t<Rng>, O>> if the execution policy is of type sequenced_task_policy or parallel_task_policy and returns util::in_out_result <traits::range_iterator_t<Rng>, O> otherwise. The exclusive_scan algorithm returns an input iterator to the point denoted by the sentinel and an output iterator to the element in the destination range, one past the last element copied.