transform_reduce.hpp¶

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

namespace hpx

Functions

template<typename ExPolicy, typename Rng, typename T, typename Reduce, typename Convert> util::detail::algorithm_result<ExPolicy, T>::type transform_reduce(ExPolicy &&policy, Rng &&rng, T init, Reduce &&red_op, Convert &&conv_op)¶

Returns GENERALIZED_SUM(red_op, init, conv_op(*first), …, conv_op(*(first + (last - first) - 1))).

The reduce operations in the parallel

transform_reduce algorithm invoked with an execution policy object of type sequenced_policy execute in sequential order in the calling thread.

Note

Complexity: O(last - first) applications of the predicates red_op and conv_op.

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 input iterator.
F: The type of the function/function object to use (deduced). Unlike its sequential form, the parallel overload of copy_if requires F to meet the requirements of CopyConstructible.
T: The type of the value to be used as initial (and intermediate) values (deduced).
Reduce: The type of the binary function object used for the reduction operation.
Convert: The type of the unary function object used to transform the elements of the input sequence before invoking the reduce function.

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.
init: The initial value for the generalized sum.
red_op: Specifies the function (or function object) which will be invoked for each of the values returned from the invocation of conv_op. This is a binary predicate. The signature of this predicate should be equivalent to:
```
Ret fun(const Type1 &a, const Type2 &b);
```
The signature does not need to have const&, but the function must not modify the objects passed to it. The types
Type1, Type2, and Ret must be such that an object of a type as returned from conv_op can be implicitly converted to any of those types.
conv_op: Specifies the function (or function object) which will be invoked for each of the elements in the sequence specified by [first, last). This is a unary predicate. The signature of this predicate should be equivalent to:
```
R fun(const Type &a);
```
The signature does not need to have const&, but the function must not modify the objects passed to it. The type
Type must be such that an object of type Iter can be dereferenced and then implicitly converted to Type. The type R must be such that an object of this type can be implicitly converted to T.

The reduce operations in the parallel transform_reduce 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

transform_reduce and accumulate is that the behavior of transform_reduce may be non-deterministic for non-associative or non-commutative binary predicate.

Return

The transform_reduce algorithm returns a hpx::future<T> if the execution policy is of type parallel_task_policy and returns T otherwise. The transform_reduce algorithm returns the result of the generalized sum over the values returned from conv_op when applied to the elements given by the input range [first, last).

Note

GENERALIZED_SUM(op, a1, …, aN) is defined as follows:

a1 when N is 1
op(GENERALIZED_SUM(op, b1, …, bK), GENERALIZED_SUM(op, bM, …, bN)), where:
- b1, …, bN may be any permutation of a1, …, aN and
- 1 < K+1 = M <= N.

template<typename ExPolicy, typename Rng1, typename FwdIter2, typename T> util::detail::algorithm_result<ExPolicy, T>::type transform_reduce(ExPolicy &&policy, Rng1 &&rng1, FwdIter2 first2, T init)¶

Returns the result of accumulating init with the inner products of the pairs formed by the elements of two ranges starting at first1 and first2.

The operations in the parallel

transform_reduce algorithm invoked with an execution policy object of type sequenced_policy execute in sequential order in the calling thread.

Note

Complexity: O(last - first) applications of the predicate op2.

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.
Rng1: The type of the source range used (deduced). The iterators extracted from this range type must meet the requirements of an input iterator.
FwdIter2: The type of the second source iterators used (deduced). This iterator type must meet the requirements of an forward iterator.
T: The type of the value to be used as return) values (deduced).

Parameters

policy: The execution policy to use for the scheduling of the iterations.
rng1: Refers to the sequence of elements the algorithm will be applied to.
first2: Refers to the beginning of the second sequence of elements the result will be calculated with.
init: The initial value for the sum.

The operations in the parallel transform_reduce 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.

Return: The transform_reduce algorithm returns a hpx::future<T> if the execution policy is of type sequenced_task_policy or parallel_task_policy and returns T otherwise.

template<typename ExPolicy, typename Rng1, typename FwdIter2, typename T, typename Reduce, typename Convert> util::detail::algorithm_result<ExPolicy, T>::type transform_reduce(ExPolicy &&policy, Rng1 &&rng1, FwdIter2 first2, T init, Reduce &&red_op, Convert &&conv_op)¶

Returns the result of accumulating init with the inner products of the pairs formed by the elements of two ranges starting at first1 and first2.

The operations in the parallel

transform_reduce algorithm invoked with an execution policy object of type sequenced_policy execute in sequential order in the calling thread.

Note

Complexity: O(last - first) applications of the predicate op2.

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.
Rng1: The type of the source range used (deduced). The iterators extracted from this range type must meet the requirements of an input iterator.
FwdIter2: The type of the second source iterators used (deduced). This iterator type must meet the requirements of an forward iterator.
T: The type of the value to be used as return) values (deduced).
Reduce: The type of the binary function object used for the multiplication operation.
Convert: The type of the unary function object used to transform the elements of the input sequence before invoking the reduce function.

Parameters

policy: The execution policy to use for the scheduling of the iterations.
rng1: Refers to the sequence of elements the algorithm will be applied to.
first2: Refers to the beginning of the second sequence of elements the result will be calculated with.
init: The initial value for the sum.
red_op: Specifies the function (or function object) which will be invoked for the initial value and each of the return values of op2. This is a binary predicate. The signature of this predicate should be equivalent to 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 type
Ret must be such that it can be implicitly converted to a type of T.
conv_op: Specifies the function (or function object) which will be invoked for each of the input values of the sequence. This is a binary predicate. The signature of this predicate should be equivalent to
```
Ret fun(const Type1 &a, const Type2 &b);
```
The signature does not need to have const&, but the function must not modify the objects passed to it. The type
Ret must be such that it can be implicitly converted to an object for the second argument type of op1.

The operations in the parallel transform_reduce 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.

Return: The transform_reduce algorithm returns a hpx::future<T> if the execution policy is of type sequenced_task_policy or parallel_task_policy and returns T otherwise.