40 : _input(nullptr), _output_x(nullptr), _output_y(nullptr), _run_sobel_x(false), _run_sobel_y(false), _border_size(0)
59 _run_sobel_x = output_x !=
nullptr;
60 _run_sobel_y = output_y !=
nullptr;
75 _border_size =
BorderSize(border_undefined ? 0 : 3, 3);
78 const std::string
kernel_name =
"sobel_separable1x7";
81 std::set<std::string> build_opts;
85 build_opts.insert(
"-DGRAD_X");
90 build_opts.insert(
"-DGRAD_Y");
94 _kernel =
create_kernel(compile_context, kernel_name, build_opts);
98 constexpr
unsigned int num_elems_read_per_iteration = 16;
99 constexpr
unsigned int num_elems_written_per_iteration = 8;
112 ICLKernel::configure_internal(win);
134 unsigned int idx = 0;
145 : _input_x(nullptr), _input_y(nullptr), _output_x(nullptr), _output_y(nullptr), _run_sobel_x(false), _run_sobel_y(false)
163 _run_sobel_x = output_x !=
nullptr;
164 _run_sobel_y = output_y !=
nullptr;
180 _output_x = output_x;
181 _output_y = output_y;
184 std::set<std::string> build_opts;
188 build_opts.insert(
"-DGRAD_X");
193 build_opts.insert(
"-DGRAD_Y");
197 const std::string
kernel_name = std::string(
"sobel_separable7x1");
198 _kernel =
create_kernel(compile_context, kernel_name, build_opts);
204 constexpr
unsigned int num_elems_written_per_iteration = 8;
205 constexpr
unsigned int num_elems_read_per_iteration = 8;
206 constexpr
unsigned int num_rows_read_per_iteration = 7;
220 ICLKernel::configure_internal(win);
243 unsigned int idx = 0;
250 _kernel.setArg(idx++, 0 );
Window first_slice_window_2D() const
First 2D slice of the window.
unsigned int top
top of the border
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
const Window & window() const
The maximum window the kernel can be executed on.
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 ...
void enqueue(IGCKernel &kernel, const Window &window, const gles::NDRange &lws=gles::NDRange(1U, 1U, 1U))
Add the kernel to the command queue with the given window.
virtual size_t dimension(size_t index) const =0
Return the size of the requested dimension.
Container for 2D border size.
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
1 channel, 1 U8 per channel
void run(const Window &window, cl::CommandQueue &queue) override
Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue...
std::string to_string(T &&value)
Convert integer and float values to string.
virtual DataType data_type() const =0
Data type used for each element of the tensor.
void run(const Window &window, cl::CommandQueue &queue) override
Enqueue the OpenCL kernel to process the given window on the passed OpenCL command queue...
#define ARM_COMPUTE_ERROR_ON(cond)
If the condition is true then an error message is printed and an exception thrown.
static CLKernelLibrary & get()
Access the KernelLibrary singleton.
void configure(const ICLTensor *input_x, const ICLTensor *input_y, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined)
Initialise the kernel's source, destination and border.
std::string lower_string(const std::string &val)
Lower a given string.
Window calculate_max_window_horizontal(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
bool slide_window_slice_2D(Window &slice) const
Slide the passed 2D window slice.
Copyright (c) 2017-2021 Arm Limited.
virtual ValidRegion valid_region() const =0
Valid region of the tensor.
CLSobel7x7VertKernel()
Default constructor: initialize all the pointers to nullptr and parameters to zero.
1 channel, 1 S32 per channel
Implementation of a rectangular access pattern.
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.
const std::string & string_from_data_type(DataType dt)
Convert a data type identity into a string.
bool update_window_and_padding(Window &win, Ts &&... patterns)
Update window and padding size for each of the access patterns.
Class to describe a number of elements in each dimension.
BorderSize border_size() const override
The size of the border for that kernel.
Implementation of a row access pattern.
virtual ITensorInfo * info() const =0
Interface to be implemented by the child class to return the tensor's metadata.
unsigned int left
left of the border
void configure(const ICLTensor *input, ICLTensor *output_x, ICLTensor *output_y, bool border_undefined)
Initialise the kernel's source, destination and border.
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
#define ARM_COMPUTE_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(t, c,...)
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...
Interface for OpenCL tensor.
Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context...
Wrapper to configure the Khronos OpenCL C++ header.
CLSobel7x7HorKernel()
Default constructor: initialize all the pointers to nullptr and parameters to zero.
unsigned int num_elems_processed_per_iteration
Describe a multidimensional execution window.
#define ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(f, s)
SimpleTensor< T > slice(const SimpleTensor< T > &src, Coordinates starts, Coordinates ends)
BorderSize border_size() const override
The size of the border for that kernel.