7.3.1. Data Structures - 5.2 English

7.3.1. Data Structures - 5.2 English - 68552

AOCL API Guide (68552)

Document ID

68552

Release Date

2025-12-29

Version

5.2 English

group Data_Structures

Data structures and typedef definitions associated with APIs of AOCL-FFTZ library.

This section contains typedef definitions used by the AOCL-FFTZ library’s interface APIs.

Note

Different variants of data structures are available to support float and double precision types in LP64 and ILP64 data models.

Typedefs

typedef struct aoclfftz_flags aoclfftz_flags_t#: Configuration flags to control critical FFT execution parameters.

typedef struct aoclfftz_dim aoclfftz_dim_t#: Tensor dimension for LP64.

typedef struct aoclfftz_dim_64_ aoclfftz_dim_t_64_#: Tensor dimension for ILP64.

typedef struct aoclfftz_smp_pfft aoclfftz_smp_pfft_t#: Params for parallel SMP FFT computation.

typedef struct aoclfftz_cntrl_params aoclfftz_cntrl_params_t#: Control params for optimizations, logs and stats.

Enums

enum aoclfftz_error_type#

Error return codes of AOCL-FFTZ library.

Values:

enumerator AOCLFFTZ_TIME_OUT#: Operation took long time than expected

enumerator AOCLFFTZ_MEMORY_FAILURE#: Error related to Memory access or operation

enumerator AOCLFFTZ_INVALID_INPUT#: Invalid size, format, type or precision of input

enumerator AOCLFFTZ_SETUP_FAILURE#: Error in setup of the library

enumerator AOCLFFTZ_EXECUTION_FAILURE#: Error in execution of the library

enumerator AOCLFFTZ_SUCCESS#: Successful operation

enum aoclfftz_logger_mode#

Logging modes to control verbosity of AOCL-FFTZ library output.

Values:

enumerator AOCLFFTZ_LOG_NONE#: 0 : Disable all logging

enumerator AOCLFFTZ_LOG_INFO#: 1 : Detailed debugging logs

enumerator AOCLFFTZ_LOG_TRACE#: 2 : Logging with complete execution trace

enumerator AOCLFFTZ_LOG_DEBUG#: 3 : Most detailed logging, with debug logs

struct aoclfftz_flags#

#include <aoclfftz.h>

Configuration flags to control critical FFT execution parameters.

Public Members

UINT8 fft_type#: Complex(0) or Real(1) >

UINT8 fft_direction#: Forward(0) or Backward(1) >

UINT8 storage_order#: In-order(0) or Out-of-order(1)>

UINT8 fft_placement#: In-place(0) or Out-of-place(1)

UINT8 transpose_mode#: fft(0) or standalone transpose(1) (Not supported, must be default value 0)

UINT8 bit_reproducibility#: Disable(0) or Enable(1) the bit reproducibility mode

struct aoclfftz_dim#

#include <aoclfftz.h>

Tensor dimension for LP64.

Public Members

INT32 n#: Dimension length

INT32 in_stride#: Stride for input buffer

INT32 out_stride#: Stride for output buffer

struct aoclfftz_dim_64_#

#include <aoclfftz.h>

Tensor dimension for ILP64.

Public Members

INTP n#: Dimension length

INTP in_stride#: Stride for input buffer

INTP out_stride#: Stride for output buffer

struct aoclfftz_smp_pfft#

#include <aoclfftz.h>

Params for parallel SMP FFT computation.

Public Members

INT32 num_threads#: Number of max threads granted for use.num_threads must be greater than 0.

UINT32 dynamic_load_model#

Specifies the model for determining the number of threads.

0: Use num_threads as the maximum number of threads.
1: Dynamically determine the number of threads (currently configured to take upto the max system threads)

struct aoclfftz_cntrl_params#

#include <aoclfftz.h>

Control params for optimizations, logs and stats.

Public Members

INT32 opt_level#

Set Optimization level with following values

Levels:

0 - non-SIMD algorithmic optimizations
1 - AVX128 optimizations
2 - AVX256 optimizations
3 - AVX512 optimizations

INT32 opt_off#: Turn off all optimizations

aoclfftz_logger_mode logger_mode#

Set Logger mode with following values.

0 - None,
1 - Info,
2 - Trace,
3 - Debug

INT32 measure_stats#: Enable/Disable measure stats (Not supported, must be default value 0)

struct aoclfftz_prob_desc_f#

#include <aoclfftz.h>

Defines problem descriptor for float data on LP64 machines.

Public Members

FLOAT *in#: input buffer: Points to signal points for forward FFT and frequency points for backward FFT

FLOAT *out#: output buffer: Points frequency points for forward FFT and signal points for backward FFT

INT32 vec_rank#: vec_rank is the number of batch/vector dimensions. length of vecs array(must be >= 1).

INT32 dim_rank#: dim_rank is the number of signal/frequency dimensions. length of dims array(must be >= 1).

aoclfftz_dim_t *dims#: Multi-dimensional tensor dimensions for LP64

aoclfftz_dim_t *vecs#: Vector tensor dimensions for LP64

aoclfftz_flags_t flags#: Struct for configuration flags to control critical FFT execution parameters like -fft_type, fft_direction, storage_order, fft_placement, transpose_mode.

aoclfftz_smp_pfft_t pthr_fft#: Struct for parallel SMP FFT computation

aoclfftz_cntrl_params_t cntrl_params#: Struct for optimizations, logs, stat params

struct aoclfftz_prob_desc_d#

#include <aoclfftz.h>

Defines problem descriptor for double data on LP64 machines.

Public Members

DOUBLE *in#: input buffer: Points to signal points for forward FFT and frequency points for backward FFT

DOUBLE *out#: output buffer: Points frequency points for forward FFT and signal points for backward FFT

INT32 vec_rank#: vec_rank is the number of batch/vector dimensions. length of vecs array(must be >= 1).

INT32 dim_rank#: dim_rank is the number of signal/frequency dimensions. length of dims array(must be >= 1).

aoclfftz_dim_t *dims#: Multi-dimensional tensor dimensions for LP64

aoclfftz_dim_t *vecs#: Vector tensor dimensions for LP64

aoclfftz_flags_t flags#: Struct for configuration flags to control critical FFT execution parameters like -fft_type, fft_direction, storage_order, fft_placement, transpose_mode.

aoclfftz_smp_pfft_t pthr_fft#: Struct for parallel SMP FFT computation

aoclfftz_cntrl_params_t cntrl_params#: Struct for optimizations, logs, stat params

struct aoclfftz_prob_desc_f_64_#

#include <aoclfftz.h>

Defines problem descriptor for float data on ILP64 machines.

Public Members

FLOAT *in#: input buffer: Points to signal points for forward FFT and frequency points for backward FFT

FLOAT *out#: output buffer: Points to frequency points for forward FFT and signal points for backward FFT

INT32 vec_rank#: vec_rank is the number of batch/vector dimensions. length of vecs array(must be >= 1).

INT32 dim_rank#: dim_rank is the number of signal/frequency dimensions. length of dims array(must be >= 1).

aoclfftz_dim_t_64_ *dims#: Multi-dimensional tensor dimensions for ILP64

aoclfftz_dim_t_64_ *vecs#: Vector tensor dimensions for ILP64

aoclfftz_flags_t flags#: Struct for configuration flags to control critical FFT execution parameters like -fft_type, fft_direction, storage_order, fft_placement, transpose_mode.

aoclfftz_smp_pfft_t pthr_fft#: Struct for parallel SMP FFT computation

aoclfftz_cntrl_params_t cntrl_params#: Struct for optimizations, logs, stat params

struct aoclfftz_prob_desc_d_64_#

#include <aoclfftz.h>

Defines problem descriptor for double data on ILP64 machines.

Public Members

DOUBLE *in#: input buffer: Points to signal points for forward FFT and frequency points for backward FFT

DOUBLE *out#: output buffer: Points to frequency points for forward FFT and signal points for backward FFT

INT32 vec_rank#: vec_rank is the number of batch/vector dimensions. length of vecs array(must be >= 1).

INT32 dim_rank#: dim_rank is the number of signal/frequency dimensions. length of dims array(must be >= 1).

aoclfftz_dim_t_64_ *dims#: Multi-dimensional tensor dimensions for ILP64

aoclfftz_dim_t_64_ *vecs#: Vector tensor dimensions for ILP64

aoclfftz_flags_t flags#: Struct for configuration flags to control critical FFT execution parameters like - fft_type, fft_direction, storage_order, fft_placement, transpose_mode.

aoclfftz_smp_pfft_t pthr_fft#: Struct for parallel SMP FFT computation

aoclfftz_cntrl_params_t cntrl_params#: Struct for optimizations, logs, stat params