cub::internal Namespace Reference

Internal namespace (to prevent ADL mishaps between static functions when mixing different CUB installations) More...

Functions
template<int LENGTH, typename T , typename ReductionOp >
__device__ __forceinline__ T	ThreadReduce (T *input, ReductionOp reduction_op, T prefix, Int2Type< LENGTH >)
template<int LENGTH, typename T , typename ReductionOp >
__device__ __forceinline__ T	ThreadReduce (T *input, ReductionOp reduction_op, T prefix)
	Perform a sequential reduction over LENGTH elements of the input array, seeded with the specified prefix. The aggregate is returned.
template<int LENGTH, typename T , typename ReductionOp >
__device__ __forceinline__ T	ThreadReduce (T *input, ReductionOp reduction_op)
	Perform a sequential reduction over LENGTH elements of the input array. The aggregate is returned.
template<int LENGTH, typename T , typename ReductionOp >
__device__ __forceinline__ T	ThreadReduce (T(&input)[LENGTH], ReductionOp reduction_op, T prefix)
	Perform a sequential reduction over the statically-sized input array, seeded with the specified prefix. The aggregate is returned.
template<int LENGTH, typename T , typename ReductionOp >
__device__ __forceinline__ T	ThreadReduce (T(&input)[LENGTH], ReductionOp reduction_op)
	Serial reduction with the specified operator.
Sequential prefix scan over statically-sized array types
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanExclusive (T inclusive, T exclusive, T *input, T *output, ScanOp scan_op, Int2Type< LENGTH >)
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanExclusive (T *input, T *output, ScanOp scan_op, T prefix, bool apply_prefix=true)
	Perform a sequential exclusive prefix scan over LENGTH elements of the input array, seeded with the specified prefix. The aggregate is returned.
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanExclusive (T(&input)[LENGTH], T(&output)[LENGTH], ScanOp scan_op, T prefix, bool apply_prefix=true)
	Perform a sequential exclusive prefix scan over the statically-sized input array, seeded with the specified prefix. The aggregate is returned.
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanInclusive (T inclusive, T *input, T *output, ScanOp scan_op, Int2Type< LENGTH >)
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanInclusive (T *input, T *output, ScanOp scan_op)
	Perform a sequential inclusive prefix scan over LENGTH elements of the input array. The aggregate is returned.
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanInclusive (T(&input)[LENGTH], T(&output)[LENGTH], ScanOp scan_op)
	Perform a sequential inclusive prefix scan over the statically-sized input array. The aggregate is returned.
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanInclusive (T *input, T *output, ScanOp scan_op, T prefix, bool apply_prefix=true)
	Perform a sequential inclusive prefix scan over LENGTH elements of the input array, seeded with the specified prefix. The aggregate is returned.
template<int LENGTH, typename T , typename ScanOp >
__device__ __forceinline__ T	ThreadScanInclusive (T(&input)[LENGTH], T(&output)[LENGTH], ScanOp scan_op, T prefix, bool apply_prefix=true)
	Perform a sequential inclusive prefix scan over the statically-sized input array, seeded with the specified prefix. The aggregate is returned.

Detailed Description

Internal namespace (to prevent ADL mishaps between static functions when mixing different CUB installations)

Function Documentation

ThreadReduce() [1/5]

template<int LENGTH, typename T , typename ReductionOp >

__device__ __forceinline__ T cub::internal::ThreadReduce	(	T *	input,
		ReductionOp	reduction_op
	)

Perform a sequential reduction over LENGTH elements of the input array. The aggregate is returned.

Template Parameters

LENGTH	LengthT of input array
T	[inferred] The data type to be reduced.
ScanOp	[inferred] Binary reduction operator type having member T operator()(const T &a, const T &b)

Parameters

[in]	input	Input array
[in]	reduction_op	Binary reduction operator

Definition at line 102 of file thread_reduce.cuh.

ThreadReduce() [2/5]

template<int LENGTH, typename T , typename ReductionOp >

__device__ __forceinline__ T cub::internal::ThreadReduce	(	T *	input,
		ReductionOp	reduction_op,
		T	prefix
	)

Perform a sequential reduction over LENGTH elements of the input array, seeded with the specified prefix. The aggregate is returned.

Template Parameters

LENGTH	LengthT of input array
T	[inferred] The data type to be reduced.
ScanOp	[inferred] Binary reduction operator type having member T operator()(const T &a, const T &b)

Parameters

[in]	input	Input array
[in]	reduction_op	Binary reduction operator
[in]	prefix	Prefix to seed reduction with

Definition at line 82 of file thread_reduce.cuh.

ThreadReduce() [3/5]

template<int LENGTH, typename T , typename ReductionOp >

__device__ __forceinline__ T cub::internal::ThreadReduce	(	T *	input,
		ReductionOp	reduction_op,
		T	prefix,
		Int2Type< LENGTH >
	)

Sequential reduction over statically-sized array types

Parameters

[in]	input	Input array
[in]	reduction_op	Binary reduction operator
[in]	prefix	Prefix to seed reduction with

Definition at line 55 of file thread_reduce.cuh.

ThreadReduce() [4/5]

template<int LENGTH, typename T , typename ReductionOp >

__device__ __forceinline__ T cub::internal::ThreadReduce	(	T(&)	input[LENGTH],
		ReductionOp	reduction_op
	)

Serial reduction with the specified operator.

Template Parameters

LENGTH	[inferred] LengthT of input array
T	[inferred] The data type to be reduced.
ScanOp	[inferred] Binary reduction operator type having member T operator()(const T &a, const T &b)

Parameters

[in]	input	Input array
[in]	reduction_op	Binary reduction operator

Definition at line 142 of file thread_reduce.cuh.

ThreadReduce() [5/5]

template<int LENGTH, typename T , typename ReductionOp >

__device__ __forceinline__ T cub::internal::ThreadReduce	(	T(&)	input[LENGTH],
		ReductionOp	reduction_op,
		T	prefix
	)

Perform a sequential reduction over the statically-sized input array, seeded with the specified prefix. The aggregate is returned.

Template Parameters

LENGTH	[inferred] LengthT of input array
T	[inferred] The data type to be reduced.
ScanOp	[inferred] Binary reduction operator type having member T operator()(const T &a, const T &b)

Parameters

[in]	input	Input array
[in]	reduction_op	Binary reduction operator
[in]	prefix	Prefix to seed reduction with

Definition at line 122 of file thread_reduce.cuh.