Compute Library
 19.08
NEFFT1D Class Reference

Basic function to execute one dimensional FFT. More...

#include <NEFFT1D.h>

Collaboration diagram for NEFFT1D:
[legend]

Public Member Functions

 NEFFT1D (std::shared_ptr< IMemoryManager > memory_manager=nullptr)
 Default Constructor. More...
 
void configure (const ITensor *input, ITensor *output, const FFT1DInfo &config)
 Initialise the function's source and destinations. More...
 
void run () override
 Run the kernels contained in the function. More...
 
- Public Member Functions inherited from IFunction
virtual ~IFunction ()=default
 Destructor. More...
 
virtual void prepare ()
 Prepare the function for executing. More...
 

Static Public Member Functions

static Status validate (const ITensorInfo *input, const ITensorInfo *output, const FFT1DInfo &config)
 Static function to check if given info will lead to a valid configuration of NEFFT1D. More...
 

Detailed Description

Basic function to execute one dimensional FFT.

This function calls the following NEON kernels:

  1. NEFFTDigitReverseKernel Performs digit reverse
  2. NEFFTRadixStageKernel A list of FFT kernels depending on the radix decomposition
  3. NEFFTScaleKernel Performs output scaling in case of in inverse FFT

Definition at line 47 of file NEFFT1D.h.

Constructor & Destructor Documentation

◆ NEFFT1D()

NEFFT1D ( std::shared_ptr< IMemoryManager memory_manager = nullptr)

Default Constructor.

Definition at line 33 of file NEFFT1D.cpp.

34  : _memory_group(std::move(memory_manager)), _digit_reverse_kernel(), _fft_kernels(), _scale_kernel(), _digit_reversed_input(), _digit_reverse_indices(), _num_ffts(0), _axis(0), _run_scale(false)
35 {
36 }

Member Function Documentation

◆ configure()

void configure ( const ITensor input,
ITensor output,
const FFT1DInfo config 
)

Initialise the function's source and destinations.

Parameters
[in]inputSource tensor. Data types supported: F32. Number of channels supported: 1 (real tensor) or 2 (complex tensor).
[out]outputDestination tensor. Data types and data layouts supported: Same as input. Number of channels supported: 1 (real tensor) or 2 (complex tensor).If input is real, output must be complex.
[in]configFFT related configuration

Definition at line 38 of file NEFFT1D.cpp.

39 {
40  ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
41  ARM_COMPUTE_ERROR_THROW_ON(NEFFT1D::validate(input->info(), output->info(), config));
42 
43  // Decompose size to radix factors
44  const auto supported_radix = NEFFTRadixStageKernel::supported_radix();
45  const unsigned int N = input->info()->tensor_shape()[config.axis];
46  const auto decomposed_vector = arm_compute::helpers::fft::decompose_stages(N, supported_radix);
47  ARM_COMPUTE_ERROR_ON(decomposed_vector.empty());
48 
49  // Flags
50  _run_scale = config.direction == FFTDirection::Inverse;
51 
52  const bool is_c2r = input->info()->num_channels() == 2 && output->info()->num_channels() == 1;
53 
54  // Configure digit reverse
55  FFTDigitReverseKernelInfo digit_reverse_config;
56  digit_reverse_config.axis = config.axis;
57  digit_reverse_config.conjugate = config.direction == FFTDirection::Inverse;
58  TensorInfo digit_reverse_indices_info(TensorShape(input->info()->tensor_shape()[config.axis]), 1, DataType::U32);
59  _digit_reverse_indices.allocator()->init(digit_reverse_indices_info);
60  _memory_group.manage(&_digit_reversed_input);
61  _digit_reverse_kernel.configure(input, &_digit_reversed_input, &_digit_reverse_indices, digit_reverse_config);
62 
63  // Create and configure FFT kernels
64  unsigned int Nx = 1;
65  _num_ffts = decomposed_vector.size();
66  _fft_kernels.resize(_num_ffts);
67  _axis = config.axis;
68 
69  for(unsigned int i = 0; i < _num_ffts; ++i)
70  {
71  const unsigned int radix_for_stage = decomposed_vector.at(i);
72 
73  FFTRadixStageKernelInfo fft_kernel_info;
74  fft_kernel_info.axis = config.axis;
75  fft_kernel_info.radix = radix_for_stage;
76  fft_kernel_info.Nx = Nx;
77  fft_kernel_info.is_first_stage = (i == 0);
78  _fft_kernels[i].configure(&_digit_reversed_input, ((i == (_num_ffts - 1)) && !is_c2r) ? output : nullptr, fft_kernel_info);
79 
80  Nx *= radix_for_stage;
81  }
82 
83  // Configure scale kernel
84  if(_run_scale)
85  {
86  FFTScaleKernelInfo scale_config;
87  scale_config.scale = static_cast<float>(N);
88  scale_config.conjugate = config.direction == FFTDirection::Inverse;
89  is_c2r ? _scale_kernel.configure(&_digit_reversed_input, output, scale_config) : _scale_kernel.configure(output, nullptr, scale_config);
90  }
91 
92  // Allocate tensors
93  _digit_reversed_input.allocator()->allocate();
94  _digit_reverse_indices.allocator()->allocate();
95 
96  // Init digit reverse indices
97  const auto digit_reverse_cpu = arm_compute::helpers::fft::digit_reverse_indices(N, decomposed_vector);
98  std::copy_n(digit_reverse_cpu.data(), N, reinterpret_cast<unsigned int *>(_digit_reverse_indices.buffer()));
99 }
void init(const TensorAllocator &allocator, const Coordinates &coords, TensorInfo &sub_info)
Shares the same backing memory with another tensor allocator, while the tensor info might be differen...
std::vector< unsigned int > decompose_stages(unsigned int N, const std::set< unsigned int > &supported_factors)
Decompose a given 1D input size using the provided supported factors.
Definition: fft.cpp:34
std::vector< unsigned int > digit_reverse_indices(unsigned int N, const std::vector< unsigned int > &fft_stages)
Calculate digit reverse index vector given fft size and the decomposed stages.
Definition: fft.cpp:79
#define ARM_COMPUTE_ERROR_ON(cond)
If the condition is true then an error message is printed and an exception thrown.
Definition: Error.h:337
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Definition: Error.h:327
TensorAllocator * allocator()
Return a pointer to the tensor's allocator.
Definition: Tensor.cpp:48
void manage(TensorType *obj)
Sets a object to be managed by the given memory group.
1 channel, 1 U32 per channel
void allocate() override
Allocate size specified by TensorInfo of CPU memory.
static std::set< unsigned int > supported_radix()
Returns the radix that are support by the FFT kernel.
void configure(ITensor *input, ITensor *output, const FFTScaleKernelInfo &config)
Set the input and output tensors.
void configure(const ITensor *input, ITensor *output, const ITensor *idx, const FFTDigitReverseKernelInfo &config)
Set the input and output tensors.
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Definition: Validate.h:161
uint8_t * buffer() const override
Interface to be implemented by the child class to return a pointer to CPU memory.
Definition: Tensor.cpp:43
static Status validate(const ITensorInfo *input, const ITensorInfo *output, const FFT1DInfo &config)
Static function to check if given info will lead to a valid configuration of NEFFT1D.
Definition: NEFFT1D.cpp:101

References TensorAllocator::allocate(), Tensor::allocator(), ARM_COMPUTE_ERROR_ON, ARM_COMPUTE_ERROR_ON_NULLPTR, ARM_COMPUTE_ERROR_THROW_ON, FFT1DInfo::axis, FFTDigitReverseKernelInfo::axis, FFTRadixStageKernelInfo::axis, Tensor::buffer(), NEFFTDigitReverseKernel::configure(), NEFFTScaleKernel::configure(), FFTScaleKernelInfo::conjugate, FFTDigitReverseKernelInfo::conjugate, arm_compute::helpers::fft::decompose_stages(), arm_compute::helpers::fft::digit_reverse_indices(), FFT1DInfo::direction, ITensor::info(), TensorAllocator::init(), arm_compute::Inverse, FFTRadixStageKernelInfo::is_first_stage, MemoryGroupBase< TensorType >::manage(), ITensorInfo::num_channels(), FFTRadixStageKernelInfo::Nx, FFTRadixStageKernelInfo::radix, FFTScaleKernelInfo::scale, NEFFTRadixStageKernel::supported_radix(), ITensorInfo::tensor_shape(), arm_compute::U32, and NEFFT1D::validate().

Referenced by NEFFT2D::configure().

◆ run()

void run ( )
overridevirtual

Run the kernels contained in the function.

For NEON kernels:

  • Multi-threading is used for the kernels which are parallelisable.
  • By default std::thread::hardware_concurrency() threads are used.
Note
CPPScheduler::set_num_threads() can be used to manually set the number of threads

For OpenCL kernels:

  • All the kernels are enqueued on the queue associated with CLScheduler.
  • The queue is then flushed.
Note
The function will not block until the kernels are executed. It is the user's responsibility to wait.
Will call prepare() on first run if hasn't been done

Implements IFunction.

Definition at line 127 of file NEFFT1D.cpp.

128 {
129  MemoryGroupResourceScope scope_mg(_memory_group);
130 
131  NEScheduler::get().schedule(&_digit_reverse_kernel, (_axis == 0 ? Window::DimY : Window::DimZ));
132 
133  for(unsigned int i = 0; i < _num_ffts; ++i)
134  {
135  NEScheduler::get().schedule(&_fft_kernels[i], (_axis == 0 ? Window::DimY : Window::DimX));
136  }
137 
138  // Run output scaling
139  if(_run_scale)
140  {
141  NEScheduler::get().schedule(&_scale_kernel, Window::DimY);
142  }
143 }
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
static constexpr size_t DimY
Alias for dimension 1 also known as Y dimension.
Definition: Window.h:45
virtual void schedule(ICPPKernel *kernel, const Hints &hints)=0
Runs the kernel in the same thread as the caller synchronously.
static constexpr size_t DimZ
Alias for dimension 2 also known as Z dimension.
Definition: Window.h:47
static IScheduler & get()
Access the scheduler singleton.
Definition: Scheduler.cpp:96

References Window::DimX, Window::DimY, Window::DimZ, Scheduler::get(), and IScheduler::schedule().

Referenced by NEFFT2D::run().

◆ validate()

Status validate ( const ITensorInfo input,
const ITensorInfo output,
const FFT1DInfo config 
)
static

Static function to check if given info will lead to a valid configuration of NEFFT1D.

Parameters
[in]inputSource tensor info. Data types supported: F32.
[in]outputDestination tensor info. Data types and data layouts supported: Same as input.
[in]configFFT related configuration
Returns
a status

Definition at line 101 of file NEFFT1D.cpp.

102 {
104  ARM_COMPUTE_RETURN_ERROR_ON(input->data_type() != DataType::F32);
105  ARM_COMPUTE_RETURN_ERROR_ON(input->num_channels() > 2);
106  ARM_COMPUTE_RETURN_ERROR_ON(std::set<unsigned int>({ 0, 1 }).count(config.axis) == 0);
107 
108  // Check if FFT is decomposable
109  const auto supported_radix = NEFFTRadixStageKernel::supported_radix();
110  const unsigned int N = input->tensor_shape()[config.axis];
111  const auto decomposed_vector = arm_compute::helpers::fft::decompose_stages(N, supported_radix);
112  ARM_COMPUTE_RETURN_ERROR_ON(decomposed_vector.empty());
113 
114  // Checks performed when output is configured
115  if((output != nullptr) && (output->total_size() != 0))
116  {
117  // All combinations are supported except real input with real output (i.e., both input channels set to 1)
118  ARM_COMPUTE_RETURN_ERROR_ON(output->num_channels() == 1 && input->num_channels() == 1);
119  ARM_COMPUTE_RETURN_ERROR_ON(output->num_channels() > 2);
122  }
123 
124  return Status{};
125 }
std::vector< unsigned int > decompose_stages(unsigned int N, const std::set< unsigned int > &supported_factors)
Decompose a given 1D input size using the provided supported factors.
Definition: fft.cpp:34
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)
Definition: Validate.h:545
1 channel, 1 F32 per channel
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
Definition: Error.h:244
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(...)
Definition: Validate.h:443
static std::set< unsigned int > supported_radix()
Returns the radix that are support by the FFT kernel.
#define ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(...)
Definition: Validate.h:163

References ARM_COMPUTE_RETURN_ERROR_ON, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES, ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES, ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR, FFT1DInfo::axis, ITensorInfo::data_type(), arm_compute::helpers::fft::decompose_stages(), arm_compute::F32, ITensorInfo::num_channels(), NEFFTRadixStageKernel::supported_radix(), ITensorInfo::tensor_shape(), and ITensorInfo::total_size().

Referenced by NEFFT1D::configure(), and NEFFT2D::validate().


The documentation for this class was generated from the following files: