CUDPP
2.1
CUDA Data-Parallel Primitives Library
|
Base class for CUDPP Plan data structures. More...
#include <cudpp_plan.h>
Inherited by CUDPPBwtPlan, CUDPPCompactPlan, CUDPPCompressPlan, CUDPPListRankPlan, CUDPPMergeSortPlan, CUDPPMtfPlan, CUDPPRadixSortPlan, CUDPPRandPlan, CUDPPReducePlan, CUDPPScanPlan, CUDPPSegmentedScanPlan, CUDPPSparseMatrixVectorMultiplyPlan, CUDPPStringSortPlan, and CUDPPTridiagonalPlan.
Public Member Functions | |
CUDPPPlan (CUDPPManager *mgr, CUDPPConfiguration config, size_t numElements, size_t numRows, size_t rowPitch) | |
Plan base class constructor. More... | |
CUDPPHandle | getHandle () |
Public Attributes | |
CUDPPConfiguration | m_config |
size_t | m_numElements |
size_t | m_numRows |
size_t | m_rowPitch |
CUDPPManager * | m_planManager |
Base class for CUDPP Plan data structures.
CUDPPPlan and its subclasses provide the internal (i.e. not visible to the library user) infrastructure for planning algorithm execution. They own intermediate storage for CUDPP algorithms as well as, in some cases, information about optimal execution configuration for the present hardware.
CUDPPPlan::CUDPPPlan | ( | CUDPPManager * | mgr, |
CUDPPConfiguration | config, | ||
size_t | numElements, | ||
size_t | numRows, | ||
size_t | rowPitch | ||
) |
Plan base class constructor.
[in] | mgr | pointer to the CUDPPManager |
[in] | config | The configuration struct specifying algorithm and options |
[in] | numElements | The maximum number of elements to be processed |
[in] | numRows | The number of rows (for 2D operations) to be processed |
[in] | rowPitch | The pitch of the rows of input data, in elements |
|
inline |
Convert this pointer to an opaque handle
CUDPPConfiguration CUDPPPlan::m_config |
Options structure
size_t CUDPPPlan::m_numElements |
Maximum number of input elements
size_t CUDPPPlan::m_numRows |
Maximum number of input rows
size_t CUDPPPlan::m_rowPitch |
Pitch of input rows in elements
CUDPPManager* CUDPPPlan::m_planManager |
pointer to the manager of this plan