CUB
|
DeviceSelect provides device-wide, parallel operations for compacting selected items from sequences of data items residing within device-accessible memory.
int32
items, where 50% of the items are randomly selected.int32
items where segments have lengths uniformly sampled from [1,1000].Definition at line 82 of file device_select.cuh.
Static Public Methods | |
template<typename InputIteratorT , typename FlagIterator , typename OutputIteratorT , typename NumSelectedIteratorT > | |
CUB_RUNTIME_FUNCTION static __forceinline__ cudaError_t | Flagged (void *d_temp_storage, size_t &temp_storage_bytes, InputIteratorT d_in, FlagIterator d_flags, OutputIteratorT d_out, NumSelectedIteratorT d_num_selected_out, int num_items, cudaStream_t stream=0, bool debug_synchronous=false) |
Uses the d_flags sequence to selectively copy the corresponding items from d_in into d_out . The total number of items selected is written to d_num_selected_out .
![]()
. | |
template<typename InputIteratorT , typename OutputIteratorT , typename NumSelectedIteratorT , typename SelectOp > | |
CUB_RUNTIME_FUNCTION static __forceinline__ cudaError_t | If (void *d_temp_storage, size_t &temp_storage_bytes, InputIteratorT d_in, OutputIteratorT d_out, NumSelectedIteratorT d_num_selected_out, int num_items, SelectOp select_op, cudaStream_t stream=0, bool debug_synchronous=false) |
Uses the select_op functor to selectively copy items from d_in into d_out . The total number of items selected is written to d_num_selected_out .
![]()
. | |
template<typename InputIteratorT , typename OutputIteratorT , typename NumSelectedIteratorT > | |
CUB_RUNTIME_FUNCTION static __forceinline__ cudaError_t | Unique (void *d_temp_storage, size_t &temp_storage_bytes, InputIteratorT d_in, OutputIteratorT d_out, NumSelectedIteratorT d_num_selected_out, int num_items, cudaStream_t stream=0, bool debug_synchronous=false) |
Given an input sequence d_in having runs of consecutive equal-valued keys, only the first key from each run is selectively copied to d_out . The total number of items selected is written to d_num_selected_out .
![]()
. | |
|
inlinestatic |
Uses the d_flags
sequence to selectively copy the corresponding items from d_in
into d_out
. The total number of items selected is written to d_num_selected_out
.
d_flags
must be castable to bool
(e.g., bool
, char
, int
, etc.).d_out
and maintain their original relative ordering.d_temp_storage
is NULL
, no work is done and the required allocation size is returned in temp_storage_bytes
.int
device vector. InputIteratorT | [inferred] Random-access input iterator type for reading input items (may be a simple pointer type) |
FlagIterator | [inferred] Random-access input iterator type for reading selection flags (may be a simple pointer type) |
OutputIteratorT | [inferred] Random-access output iterator type for writing selected items (may be a simple pointer type) |
NumSelectedIteratorT | [inferred] Output iterator type for recording the number of items selected (may be a simple pointer type) |
[in] | d_temp_storage | Device-accessible allocation of temporary storage. When NULL, the required allocation size is written to temp_storage_bytes and no work is done. |
[in,out] | temp_storage_bytes | Reference to size in bytes of d_temp_storage allocation |
[in] | d_in | Pointer to the input sequence of data items |
[in] | d_flags | Pointer to the input sequence of selection flags |
[out] | d_out | Pointer to the output sequence of selected data items |
[out] | d_num_selected_out | Pointer to the output total number of items selected (i.e., length of d_out ) |
[in] | num_items | Total number of input items (i.e., length of d_in ) |
[in] | stream | [optional] CUDA stream to launch kernels within. Default is stream0. |
[in] | debug_synchronous | [optional] Whether or not to synchronize the stream after every kernel launch to check for errors. May cause significant slowdown. Default is false . |
Definition at line 133 of file device_select.cuh.
|
inlinestatic |
Uses the select_op
functor to selectively copy items from d_in
into d_out
. The total number of items selected is written to d_num_selected_out
.
d_out
and maintain their original relative ordering.d_temp_storage
is NULL
, no work is done and the required allocation size is returned in temp_storage_bytes
.int32
and int64
items, respectively. Items are selected with 50% probability.int
device vector. InputIteratorT | [inferred] Random-access input iterator type for reading input items (may be a simple pointer type) |
OutputIteratorT | [inferred] Random-access output iterator type for writing selected items (may be a simple pointer type) |
NumSelectedIteratorT | [inferred] Output iterator type for recording the number of items selected (may be a simple pointer type) |
SelectOp | [inferred] Selection operator type having member bool operator()(const T &a) |
[in] | d_temp_storage | Device-accessible allocation of temporary storage. When NULL, the required allocation size is written to temp_storage_bytes and no work is done. |
[in,out] | temp_storage_bytes | Reference to size in bytes of d_temp_storage allocation |
[in] | d_in | Pointer to the input sequence of data items |
[out] | d_out | Pointer to the output sequence of selected data items |
[out] | d_num_selected_out | Pointer to the output total number of items selected (i.e., length of d_out ) |
[in] | num_items | Total number of input items (i.e., length of d_in ) |
[in] | select_op | Unary selection operator |
[in] | stream | [optional] CUDA stream to launch kernels within. Default is stream0. |
[in] | debug_synchronous | [optional] Whether or not to synchronize the stream after every kernel launch to check for errors. May cause significant slowdown. Default is false . |
Definition at line 239 of file device_select.cuh.
|
inlinestatic |
Given an input sequence d_in
having runs of consecutive equal-valued keys, only the first key from each run is selectively copied to d_out
. The total number of items selected is written to d_num_selected_out
.
==
equality operator is used to determine whether keys are equivalentd_out
and maintain their original relative ordering.d_temp_storage
is NULL
, no work is done and the required allocation size is returned in temp_storage_bytes
.int32
and int64
items, respectively. Segments have lengths uniformly sampled from [1,1000].int
device vector. InputIteratorT | [inferred] Random-access input iterator type for reading input items (may be a simple pointer type) |
OutputIteratorT | [inferred] Random-access output iterator type for writing selected items (may be a simple pointer type) |
NumSelectedIteratorT | [inferred] Output iterator type for recording the number of items selected (may be a simple pointer type) |
[in] | d_temp_storage | Device-accessible allocation of temporary storage. When NULL, the required allocation size is written to temp_storage_bytes and no work is done. |
[in,out] | temp_storage_bytes | Reference to size in bytes of d_temp_storage allocation |
[in] | d_in | Pointer to the input sequence of data items |
[out] | d_out | Pointer to the output sequence of selected data items |
[out] | d_num_selected_out | Pointer to the output total number of items selected (i.e., length of d_out ) |
[in] | num_items | Total number of input items (i.e., length of d_in ) |
[in] | stream | [optional] CUDA stream to launch kernels within. Default is stream0. |
[in] | debug_synchronous | [optional] Whether or not to synchronize the stream after every kernel launch to check for errors. May cause significant slowdown. Default is false . |
Definition at line 329 of file device_select.cuh.