Compute Library
 23.11
Public Member Functions | Static Public Member Functions
ClGemmLowpOffsetContributionKernel Class Reference

OpenCL kernel used to add the offset contribution after the matrix multiplication. More...

#include <ClGemmLowpOffsetContributionKernel.h>

Collaboration diagram for ClGemmLowpOffsetContributionKernel:
[legend]

Public Member Functions

 ClGemmLowpOffsetContributionKernel ()
 
 ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE (ClGemmLowpOffsetContributionKernel)
 
void configure (const CLCompileContext &compile_context, const ITensorInfo *mm_result, const ITensorInfo *vector_sum_col, const ITensorInfo *vector_sum_row, const ITensorInfo *bias, int32_t k, int32_t a_offset, int32_t b_offset)
 Initialise the kernel's input and output. More...
 
void run_op (ITensorPack &tensors, const Window &window, cl::CommandQueue &queue) override
 Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue. More...
 
- Public Member Functions inherited from ICLKernel
 ICLKernel ()
 Constructor. More...
 
cl::Kernel & kernel ()
 Returns a reference to the OpenCL kernel of this object. More...
 
CLKernelType type () const
 Returns the CL kernel type. More...
 
template<typename T >
void add_1D_array_argument (unsigned int &idx, const ICLArray< T > *array, const Strides &strides, unsigned int num_dimensions, const Window &window)
 Add the passed 1D array's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_1D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_1D_tensor_argument_if (bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx if the condition is true. More...
 
void add_2D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_2D_tensor_argument_if (bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx if the condition is true. More...
 
void add_3D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 3D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_4D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 4D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_5D_tensor_argument (unsigned int &idx, const ICLTensor *tensor, const Window &window)
 Add the passed 5D tensor's parameters to the object's kernel's arguments starting from the index idx. More...
 
void add_3d_tensor_nhw_argument (unsigned int &idx, const ICLTensor *tensor)
 Add the passed NHW 3D tensor's parameters to the object's kernel's arguments by passing strides, dimensions and the offset to the first valid element in bytes. More...
 
void add_4d_tensor_nhwc_argument (unsigned int &idx, const ICLTensor *tensor)
 Add the passed NHWC 4D tensor's parameters to the object's kernel's arguments by passing strides, dimensions and the offset to the first valid element in bytes. More...
 
virtual void run (const Window &window, cl::CommandQueue &queue)
 Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue. More...
 
template<typename T >
void add_argument (unsigned int &idx, T value)
 Add the passed parameters to the object's kernel's arguments starting from the index idx. More...
 
void set_lws_hint (const cl::NDRange &lws_hint)
 Set the Local-Workgroup-Size hint. More...
 
cl::NDRange lws_hint () const
 Return the Local-Workgroup-Size hint. More...
 
void set_wbsm_hint (const cl_int &wbsm_hint)
 Set the workgroup batch size modifier hint. More...
 
cl_int wbsm_hint () const
 Return the workgroup batch size modifier hint. More...
 
const std::string & config_id () const
 Get the configuration ID. More...
 
void set_target (GPUTarget target)
 Set the targeted GPU architecture. More...
 
void set_target (cl::Device &device)
 Set the targeted GPU architecture according to the CL device. More...
 
GPUTarget get_target () const
 Get the targeted GPU architecture. More...
 
size_t get_max_workgroup_size ()
 Get the maximum workgroup size for the device the CLKernelLibrary uses. More...
 
cl::NDRange get_cached_gws () const
 Get the cached gws used to enqueue this kernel. More...
 
void cache_gws (const cl::NDRange &gws)
 Cache the latest gws used to enqueue this kernel. More...
 
template<unsigned int dimension_size>
void add_tensor_argument (unsigned &idx, const ICLTensor *tensor, const Window &window)
 
template<typename T , unsigned int dimension_size>
void add_array_argument (unsigned &idx, const ICLArray< T > *array, const Strides &strides, unsigned int num_dimensions, const Window &window)
 Add the passed array's parameters to the object's kernel's arguments starting from the index idx. More...
 
- Public Member Functions inherited from IKernel
 IKernel ()
 Constructor. More...
 
virtual ~IKernel ()=default
 Destructor. More...
 
virtual bool is_parallelisable () const
 Indicates whether or not the kernel is parallelisable. More...
 
virtual BorderSize border_size () const
 The size of the border for that kernel. More...
 
const Window & window () const
 The maximum window the kernel can be executed on. More...
 
bool is_window_configured () const
 Function to check if the embedded window of this kernel has been configured. More...
 

Static Public Member Functions

static Status validate (const ITensorInfo *mm_result, const ITensorInfo *vector_sum_col, const ITensorInfo *vector_sum_row, const ITensorInfo *bias, int32_t a_offset, int32_t b_offset)
 Static function to check if given info will lead to a valid configuration. More...
 
- Static Public Member Functions inherited from ICLKernel
constexpr static unsigned int num_arguments_per_3d_tensor_nhw ()
 Returns the number of arguments enqueued per NHW 3D Tensor object. More...
 
constexpr static unsigned int num_arguments_per_4d_tensor_nhwc ()
 Returns the number of arguments enqueued per NHWC 4D Tensor object. More...
 
constexpr static unsigned int num_arguments_per_1D_array ()
 Returns the number of arguments enqueued per 1D array object. More...
 
constexpr static unsigned int num_arguments_per_1D_tensor ()
 Returns the number of arguments enqueued per 1D tensor object. More...
 
constexpr static unsigned int num_arguments_per_2D_tensor ()
 Returns the number of arguments enqueued per 2D tensor object. More...
 
constexpr static unsigned int num_arguments_per_3D_tensor ()
 Returns the number of arguments enqueued per 3D tensor object. More...
 
constexpr static unsigned int num_arguments_per_4D_tensor ()
 Returns the number of arguments enqueued per 4D tensor object. More...
 
static cl::NDRange gws_from_window (const Window &window, bool use_dummy_work_items)
 Get the global work size given an execution window. More...
 

Detailed Description

OpenCL kernel used to add the offset contribution after the matrix multiplication.

The computation is performed in-place

This kernel takes a final int32 accumulator value (the output of the matrix multiplication), and adds to it the offset contribution of matrix A and matrix B in-place.

The final result is:

mm_result[i][k] = mm_result[i][k] + (vector_sum_col[k] * a_offset) + (vector_sum_row[i] * b_offset) + (a_offset * b_offset * k)

Definition at line 50 of file ClGemmLowpOffsetContributionKernel.h.

Constructor & Destructor Documentation

◆ ClGemmLowpOffsetContributionKernel()

ClGemmLowpOffsetContributionKernel ( )

Definition at line 112 of file ClGemmLowpOffsetContributionKernel.cpp.

113 {
114  _type = CLKernelType::ELEMENTWISE;
115 }

References arm_compute::ELEMENTWISE.

Member Function Documentation

◆ ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE()

ARM_COMPUTE_DISALLOW_COPY_ALLOW_MOVE ( ClGemmLowpOffsetContributionKernel  )

◆ configure()

void configure ( const CLCompileContext &  compile_context,
const ITensorInfo *  mm_result,
const ITensorInfo *  vector_sum_col,
const ITensorInfo *  vector_sum_row,
const ITensorInfo *  bias,
int32_t  k,
int32_t  a_offset,
int32_t  b_offset 
)

Initialise the kernel's input and output.

Parameters
[in]compile_contextThe compile context to be used.
[in,out]mm_resultInput tensor containing the result of the matrix multiplication. Data type supported: S32
[in]vector_sum_colInput row-vector of sums of all the entries in each column of matrix B. Note: vector_sum_col can be a nullptr in case a_offset = 0. Data type supported: same as mm_result
[in]vector_sum_rowInput row-vector of sums of all the entries in each row of matrix A. Note: vector_sum_row can be a nullptr in case b_offset = 0. Data type supported: same as mm_result
[in]biasBiases tensor. Only shared biases supported and it can be a nullptr if the addition of biases is not required. Biases are 1D tensor with dimensions [OFM]. Data type supported: Same as input.
[in]kNumber of matrix A columns or Matrix B rows
[in]a_offsetOffset to be added to each element of the matrix A.
[in]b_offsetOffset to be added to each element of the matrix B.

Definition at line 117 of file ClGemmLowpOffsetContributionKernel.cpp.

125 {
126  // Perform validate step
127  ARM_COMPUTE_ERROR_ON_NULLPTR(mm_result);
128  ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(mm_result, vector_sum_col, vector_sum_row, bias, a_offset, b_offset));
129 
130  auto padding_info = get_padding_info({mm_result, vector_sum_col, vector_sum_row, bias});
131 
132  // Check if input is a 3D reinterpretation
133  const bool reinterpret_as_3d = vector_sum_row != nullptr && mm_result->num_dimensions() > 1 &&
134  mm_result->tensor_shape().y() != vector_sum_row->tensor_shape().x();
135 
136  const unsigned int num_elems_processed_per_iteration = adjust_vec_size(4, mm_result->dimension(0));
137 
138  // Set the arguments to pass at compile time
139  CLBuildOptions build_opts;
140  build_opts.add_option("-DVEC_SIZE=" + support::cpp11::to_string(num_elems_processed_per_iteration));
141  build_opts.add_option("-DVEC_SIZE_LEFTOVER=" +
142  support::cpp11::to_string(mm_result->dimension(0) % num_elems_processed_per_iteration));
143 
144  // If a_offset == 0, vector_sum_col can be a nullptr
145  if (a_offset != 0)
146  {
147  build_opts.add_option("-DA_OFFSET=" + support::cpp11::to_string(a_offset));
148  build_opts.add_option_if(vector_sum_col->tensor_shape().num_dimensions() > 1, "-DSUM_COL_HAS_BATCHES");
149  }
150  // If b_offset == 0, vector_sum_row can be a nullptr
151  build_opts.add_option_if(b_offset != 0, "-DB_OFFSET=" + support::cpp11::to_string(b_offset));
152  build_opts.add_option("-DK_OFFSET=" + support::cpp11::to_string(a_offset * b_offset * k));
153  build_opts.add_option_if(reinterpret_as_3d,
154  "-DHEIGHT_INPUT3D=" + support::cpp11::to_string(mm_result->dimension(1)));
155  build_opts.add_option_if(reinterpret_as_3d,
156  "-DDEPTH_INPUT3D=" + support::cpp11::to_string(mm_result->dimension(2)));
157  build_opts.add_option_if(bias != nullptr, "-DADD_BIAS");
158 
159  std::string kernel_name("gemmlowp_offset_contribution");
160 
161  // A macro guard to compile ONLY the kernel of interest
162  build_opts.add_option("-D" + upper_string(kernel_name));
163 
164  // Create kernel
165  _kernel = create_kernel(compile_context, kernel_name, build_opts.options());
166 
167  // Configure kernel window
168  Window win = calculate_max_window(*mm_result, Steps(num_elems_processed_per_iteration));
169  IClKernel::configure_internal(win);
170 
171  // Set config_id for enabling LWS tuning
172  _config_id = kernel_name + "_";
173  _config_id += support::cpp11::to_string(mm_result->dimension(0));
174  _config_id += "_";
175  _config_id += support::cpp11::to_string(mm_result->dimension(1));
176  _config_id += "_";
177  _config_id += support::cpp11::to_string(mm_result->dimension(2));
178 
179  ARM_COMPUTE_ERROR_ON(has_padding_changed(padding_info));
180 }

References CLBuildOptions::add_option(), CLBuildOptions::add_option_if(), arm_compute::adjust_vec_size(), ARM_COMPUTE_ERROR_ON, ARM_COMPUTE_ERROR_ON_NULLPTR, ARM_COMPUTE_ERROR_THROW_ON, bias, arm_compute::calculate_max_window(), arm_compute::create_kernel(), ITensorInfo::dimension(), arm_compute::get_padding_info(), arm_compute::has_padding_changed(), kernel_name, Dimensions< T >::num_dimensions(), ITensorInfo::num_dimensions(), num_elems_processed_per_iteration, CLBuildOptions::options(), ITensorInfo::tensor_shape(), arm_compute::support::cpp11::to_string(), arm_compute::upper_string(), arm_compute::cpu::kernels::validate_arguments(), Dimensions< T >::x(), and Dimensions< T >::y().

◆ run_op()

void run_op ( ITensorPack &  tensors,
const Window &  window,
cl::CommandQueue &  queue 
)
overridevirtual

Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue.

Note
The queue is not flushed by this method, and therefore the kernel will not have been executed by the time this method returns.
Parameters
[in]tensorsA vector containing the tensors to operato on.
[in]windowRegion on which to execute the kernel. (Must be a valid region of the window returned by window()).
[in,out]queueCommand queue on which to enqueue the kernel.

Reimplemented from ICLKernel.

Definition at line 194 of file ClGemmLowpOffsetContributionKernel.cpp.

195 {
196  ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
197  ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(IClKernel::window(), window);
198 
199  const auto vector_sum_col =
200  utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_VEC_COL_SUM));
201  const auto vector_sum_row =
202  utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_VEC_ROW_SUM));
203  const auto bias =
204  utils::cast::polymorphic_downcast<const ICLTensor *>(tensors.get_const_tensor(TensorType::ACL_BIAS));
205  const auto mm_result = utils::cast::polymorphic_downcast<ICLTensor *>(tensors.get_tensor(TensorType::ACL_SRC_DST));
206 
207  Window collapsed = window.collapse_if_possible(IClKernel::window(), Window::DimZ);
208  Window slice = collapsed.first_slice_window_3D();
209 
210  // Set window for vector_sum_col
211  Window win_vector_sum_col = slice;
212  win_vector_sum_col.set(Window::DimY, Window::Dimension(0, 0, 0));
213  win_vector_sum_col.set(Window::DimZ, Window::Dimension(0, 0, 0));
214 
215  // Set window for vector_sum_row
216  Window win_vector_sum_row = slice;
217  win_vector_sum_row.set(Window::DimX, Window::Dimension(0, 0, 0));
218  win_vector_sum_row.set(Window::DimY, Window::Dimension(0, 0, 0));
219  win_vector_sum_col.set(Window::DimZ, Window::Dimension(0, 0, 0));
220 
221  Window biases_slice = slice;
222  biases_slice.set(Window::DimY, Window::Dimension(0, 1, 1));
223  biases_slice.set(Window::DimZ, Window::Dimension(0, 1, 1));
224 
225  do
226  {
227  unsigned int idx = 0;
228  add_3D_tensor_argument(idx, mm_result, slice);
229  add_2D_tensor_argument_if((vector_sum_col != nullptr), idx, vector_sum_col, win_vector_sum_col);
230  add_2D_tensor_argument_if((vector_sum_row != nullptr), idx, vector_sum_row, win_vector_sum_row);
231  add_1D_tensor_argument_if((bias != nullptr), idx, bias, biases_slice);
232 
233  enqueue(queue, *this, slice, lws_hint());
234  } while (collapsed.slide_window_slice_3D(slice));
235 }

References arm_compute::ACL_BIAS, arm_compute::ACL_SRC_DST, arm_compute::ACL_VEC_COL_SUM, arm_compute::ACL_VEC_ROW_SUM, ICLKernel::add_1D_tensor_argument_if(), ICLKernel::add_2D_tensor_argument_if(), ICLKernel::add_3D_tensor_argument(), ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW, ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL, bias, Window::collapse_if_possible(), Window::DimX, Window::DimY, Window::DimZ, arm_compute::enqueue(), Window::first_slice_window_3D(), ITensorPack::get_const_tensor(), ITensorPack::get_tensor(), ICLKernel::lws_hint(), Window::set(), arm_compute::test::validation::reference::slice(), Window::slide_window_slice_3D(), and IKernel::window().

◆ validate()

Status validate ( const ITensorInfo *  mm_result,
const ITensorInfo *  vector_sum_col,
const ITensorInfo *  vector_sum_row,
const ITensorInfo *  bias,
int32_t  a_offset,
int32_t  b_offset 
)
static

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

Similar to ClGemmLowpOffsetContributionKernel::configure()

Returns
a status

Definition at line 182 of file ClGemmLowpOffsetContributionKernel.cpp.

188 {
189  ARM_COMPUTE_RETURN_ON_ERROR(
190  validate_arguments(mm_result, vector_sum_col, vector_sum_row, bias, a_offset, b_offset));
191  return Status{};
192 }

References ARM_COMPUTE_RETURN_ON_ERROR, bias, and arm_compute::cpu::kernels::validate_arguments().

Referenced by ClGemmLowpMatrixMultiplyCore::validate().

The documentation for this class was generated from the following files:
arm_compute::support::cpp11::to_string
std::string to_string(T &&value)
Convert integer and float values to string.
Definition: StringSupport.h:168
arm_compute::ACL_SRC_DST
@ ACL_SRC_DST
Definition: Types.h:41
arm_compute::calculate_max_window
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
Definition: WindowHelpers.cpp:29
arm_compute::cpu::kernels::validate_arguments
Status validate_arguments(const ITensorInfo *src, const ITensorInfo *weights, const ITensorInfo *dst, const PadStrideInfo &conv_info)
Definition: CpuDirectConv2dKernel.cpp:57
ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
Definition: Validate.h:1079
arm_compute::Window::DimX
static constexpr size_t DimX
Alias for dimension 0 also known as X dimension.
Definition: Window.h:43
arm_compute::Window::collapse_if_possible
Window collapse_if_possible(const Window &full_window, size_t first, size_t last, bool *has_collapsed=nullptr) const
Collapse the dimensions between first and last if possible.
Definition: Window.inl:72
arm_compute::upper_string
std::string upper_string(const std::string &val)
Raise a given string to upper case.
Definition: StringUtils.cpp:45
ARM_COMPUTE_RETURN_ON_ERROR
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
Definition: Error.h:205
ARM_COMPUTE_ERROR_ON_NULLPTR
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Definition: Validate.h:159
ARM_COMPUTE_ERROR_ON
#define ARM_COMPUTE_ERROR_ON(cond)
If the condition is true then an error message is printed and an exception thrown.
Definition: Error.h:466
arm_compute::ICLKernel::add_2D_tensor_argument_if
void add_2D_tensor_argument_if(bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
Add the passed 2D tensor's parameters to the object's kernel's arguments starting from the index idx ...
Definition: ICLKernel.h:221
ARM_COMPUTE_ERROR_THROW_ON
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Definition: Error.h:455
arm_compute::create_kernel
cl::Kernel create_kernel(const CLCompileContext &ctx, const std::string &kernel_name, const std::set< std::string > &build_opts=std::set< std::string >())
Creates an opencl kernel using a compile context.
Definition: CLHelpers.cpp:409
ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
Definition: Validate.h:203
bias
const int32_t * bias
Definition: working_space.hpp:322
arm_compute::Window::DimY
static constexpr size_t DimY
Alias for dimension 1 also known as Y dimension.
Definition: Window.h:45
arm_compute::IKernel::window
const Window & window() const
The maximum window the kernel can be executed on.
Definition: IKernel.cpp:28
arm_compute::ICLKernel::add_3D_tensor_argument
void add_3D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window)
Add the passed 3D tensor's parameters to the object's kernel's arguments starting from the index idx.
Definition: ICLKernel.h:234
arm_compute::ICLKernel::add_1D_tensor_argument_if
void add_1D_tensor_argument_if(bool cond, unsigned int &idx, const ICLTensor *tensor, const Window &window)
Add the passed 1D tensor's parameters to the object's kernel's arguments starting from the index idx ...
Definition: ICLKernel.h:197
arm_compute::ELEMENTWISE
@ ELEMENTWISE
Elementwise CL kernel type.
Definition: CLTypes.h:83
arm_compute::ACL_BIAS
@ ACL_BIAS
Definition: Types.h:74
arm_compute::ACL_VEC_COL_SUM
@ ACL_VEC_COL_SUM
Definition: Types.h:78
arm_compute::adjust_vec_size
unsigned int adjust_vec_size(unsigned int vec_size, size_t dim0)
Returns the adjusted vector size in case it is less than the input's first dimension,...
Definition: AdjustVecSize.h:38
arm_compute::has_padding_changed
bool has_padding_changed(const std::unordered_map< const ITensorInfo *, PaddingSize > &padding_map)
Check if the previously stored padding info has changed after configuring a kernel.
Definition: Utils.cpp:491
arm_compute::Window::DimZ
static constexpr size_t DimZ
Alias for dimension 2 also known as Z dimension.
Definition: Window.h:47
arm_compute::ICLKernel::lws_hint
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
Definition: ICLKernel.h:383
num_elems_processed_per_iteration
unsigned int num_elems_processed_per_iteration
Definition: ClIm2ColKernel.cpp:60
arm_compute::get_padding_info
std::unordered_map< const ITensorInfo *, PaddingSize > get_padding_info(std::initializer_list< const ITensorInfo * > infos)
Stores padding information before configuring a kernel.
Definition: Utils.cpp:476
arm_compute::ACL_VEC_ROW_SUM
@ ACL_VEC_ROW_SUM
Definition: Types.h:77
arm_compute::test::validation::reference::slice
SimpleTensor< T > slice(const SimpleTensor< T > &src, Coordinates starts, Coordinates ends)
Definition: SliceOperations.cpp:38
kernel_name
std::string kernel_name
Definition: ClIm2ColKernel.cpp:58
arm_compute::enqueue
void enqueue(cl::CommandQueue &queue, ICLKernel &kernel, const Window &window, const cl::NDRange &lws_hint=CLKernelLibrary::get().default_ndrange(), bool use_dummy_work_items=false)
Add the kernel to the command queue with the given window.
Definition: ICLKernel.cpp:33