NFFT  3.5.3
Public Member Functions | Data Fields
nfftf_plan Struct Reference

data structure for an NFFT (nonequispaced fast Fourier transform) plan with float precision More...

#include <nfft3.h>

Public Member Functions

 FFTW_MANGLE_FLOAT (plan) my_fftw_plan1
 Forward FFTW plan.
 
 FFTW_MANGLE_FLOAT (plan) my_fftw_plan2
 Backward FFTW plan.
 

Data Fields

NFFT_INT N_total
 Total number of Fourier coefficients.
 
NFFT_INT M_total
 Total number of samples.
 
fftwf_complex * f_hat
 Fourier coefficients.
 
fftwf_complex * f
 Samples.
 
void(* mv_trafo )(void *)
 Transform.
 
void(* mv_adjoint )(void *)
 Adjoint transform.
 
NFFT_INT d
 Dimension (rank).
 
NFFT_INT * N
 Multi-bandwidth.
 
float * sigma
 Oversampling factor.
 
NFFT_INT * n
 Length of FFTW transforms. More...
 
NFFT_INT n_total
 Combined total length of FFTW transforms.
 
NFFT_INT m
 Cut-off parameter for window function. More...
 
float * b
 Shape parameter for window function.
 
NFFT_INT K
 Number of equispaced samples of window function. More...
 
unsigned flags
 Flags for precomputation, (de)allocation, and FFTW usage, default setting is PRE_PHI_HUT | PRE_PSI | MALLOC_X | MALLOC_F_HAT | MALLOC_F | FFTW_INIT | FFT_OUT_OF_PLACE.
 
unsigned fftw_flags
 Flags for the FFTW, default is FFTW_ESTIMATE | FFTW_DESTROY_INPUT.
 
float * x
 Nodes in time/spatial domain, size is $dM$ floats.
 
float MEASURE_TIME_t [3]
 Measured time for each step if MEASURE_TIME is set.
 
float ** c_phi_inv
 Precomputed data for the diagonal matrix $D$, size is $N_0+\dots+N_{d-1}$ doubles.
 
float * psi
 Precomputed data for the sparse matrix $B$, size depends on precomputation scheme.
 
NFFT_INT * psi_index_g
 Indices in source/target vector for PRE_FULL_PSI.
 
NFFT_INT * psi_index_f
 Indices in source/target vector for PRE_FULL_PSI.
 
fftwf_complex * g
 Oversampled vector of samples, size is n_total double complex.
 
fftwf_complex * g_hat
 Zero-padded vector of Fourier coefficients, size is n_total fftw_complex.
 
fftwf_complex * g1
 Input of fftw.
 
fftwf_complex * g2
 Output of fftw.
 
float * spline_coeffs
 Input for de Boor algorithm if B_SPLINE or SINC_POWER is defined.
 
NFFT_INT * index_x
 Index array for nodes x used when flag NFFT_SORT_NODES is set.
 

Detailed Description

data structure for an NFFT (nonequispaced fast Fourier transform) plan with float precision

Definition at line 179 of file nfft3.h.

Field Documentation

◆ n

NFFT_INT* nfftf_plan::n

Length of FFTW transforms.

This is equal to sigma*N. The default is to use a power of two that satifies $2\le\sigma<4$.

Definition at line 179 of file nfft3.h.

◆ m

NFFT_INT nfftf_plan::m

Cut-off parameter for window function.

Default values for the different window functions are - 6 (KAISER_BESSEL), - 9 (SINC_POWER), - 11 (B_SPLINE), - 12 (GAUSSIAN)

Definition at line 179 of file nfft3.h.

◆ K

NFFT_INT nfftf_plan::K

Number of equispaced samples of window function.

Used for flag PRE_LIN_PSI.

Definition at line 179 of file nfft3.h.


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