heffte::plan_cufft< scalar_type > Struct Template Reference

Wrapper around cufftHandle plans, set for float or double complex. More...

#include <heffte_backend_cuda.h>

Public Member Functions
	plan_cufft (cudaStream_t stream, int size, int batch, int stride, int dist)
	Constructor, takes inputs identical to cufftMakePlanMany().
	plan_cufft (cudaStream_t stream, int size1, int size2, std::array< int, 2 > const &embed, int batch, int stride, int dist)
	Constructor, takes inputs identical to cufftMakePlanMany().
	plan_cufft (cudaStream_t stream, int size1, int size2, int size3)
	Constructor, takes inputs identical to cufftPlan3d().
	~plan_cufft ()
	Destructor, deletes the plan.
	operator cufftHandle () const
	Custom conversion to the cufftHandle.

Detailed Description

template<typename scalar_type>
struct heffte::plan_cufft< scalar_type >

Wrapper around cufftHandle plans, set for float or double complex.

• Template Parameters

  scalar_type

  must be std::compelx<float> or std::complex<double>

Constructor & Destructor Documentation

plan_cufft() [1/2]

template<typename scalar_type>

heffte::plan_cufft< scalar_type >::plan_cufft ( cudaStream_t stream, int size, int batch, int stride, int dist )

inline

Constructor, takes inputs identical to cufftMakePlanMany().

Parameters

stream	is the CUDA stream to use for the transform
size	is the number of entries in a 1-D transform
batch	is the number of transforms in the batch
stride	is the distance between entries of the same transform
dist	is the distance between the first entries of consecutive sequences

plan_cufft() [2/2]

template<typename scalar_type>

heffte::plan_cufft< scalar_type >::plan_cufft ( cudaStream_t stream, int size1, int size2, std::array< int, 2 > const & embed, int batch, int stride, int dist )

inline

Constructor, takes inputs identical to cufftMakePlanMany().

Parameters

stream	is the CUDA stream to use for the transform
size1	is the number of entries in a 2-D transform (dimension 1)
size2	is the number of entries in a 2-D transform (dimension 2)
embed	defines the embedding of the 2-D transform into a 3-D box
batch	is the number of transforms in the batch
stride	is the distance between entries of the same transform
dist	is the distance between the first entries of consecutive sequences

---

The documentation for this struct was generated from the following file:

• include/heffte_backend_cuda.h