50 Status validate_arguments(
const ITensorInfo *
src,
const ITensorInfo *
dst,
const GEMMRHSMatrixInfo &rhs_info)
56 ARM_COMPUTE_RETURN_ERROR_ON_MSG(((rhs_info.k0 & (rhs_info.k0 - 1)) && (rhs_info.k0 != 1) && (rhs_info.k0 != 3)),
"Only 1,2,3,4,8,16 are supported for k0");
64 if(rhs_info.export_to_cl_image)
70 if(dst->total_size() != 0)
80 std::pair<Status, Window> validate_and_configure_window(ITensorInfo *src, ITensorInfo *dst,
const GEMMRHSMatrixInfo &rhs_info)
82 const unsigned int num_elems_processed_per_iteration_x = rhs_info.n0;
83 const unsigned int num_elems_processed_per_iteration_y = rhs_info.k0;
84 bool window_changed =
false;
90 Window win =
calculate_max_window(*src, Steps(num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y));
92 AccessWindowRectangle src_access(src, 0, 0, num_elems_processed_per_iteration_x, num_elems_processed_per_iteration_y);
96 if(rhs_info.export_to_cl_image)
106 return std::make_pair(err, collapsed);
132 std::string
kernel_name(
"gemm_reshape_rhs_matrix_");
133 kernel_name += rhs_info.
transpose ?
"t" :
"nt";
139 auto win_config = validate_and_configure_window(src, dst, rhs_info);
141 ICLKernel::configure_internal(win_config.second);
166 unsigned int idx = 0;
Window calculate_max_window(const ValidRegion &valid_region, const Steps &steps, bool skip_border, BorderSize border_size)
#define ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(tensor)
const Window & window() const
The maximum window the kernel can be executed on.
virtual size_t dimension(size_t index) const =0
Return the size of the requested dimension.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_QUANTIZATION_INFO(...)
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.
const StringSet & options() const
Gets the current options list set.
cl::NDRange lws_hint() const
Return the Local-Workgroup-Size hint.
#define ARM_COMPUTE_RETURN_ON_ERROR(status)
Checks if a status contains an error and returns it.
std::string to_string(T &&value)
Convert integer and float values to string.
unsigned int h0
Number of horizontal blocks of size (k0xn0) stored on the same output row.
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...
Store the tensor's metadata.
#define ARM_COMPUTE_ERROR_THROW_ON(status)
Manages all the OpenCL kernels compilation and caching, provides accessors for the OpenCL Context...
#define ARM_COMPUTE_RETURN_ERROR_ON(cond)
If the condition is true, an error is returned.
void configure(const ClCompileContext &compile_context, ITensorInfo *src, ITensorInfo *dst, const GEMMRHSMatrixInfo &rhs_info)
Initialise the kernel's input and output.
void update_padding_for_cl_image(ITensorInfo *tensor)
Update padding required to export the OpenCL buffer to OpenCL image2d.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(...)
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...
bool transpose
True if the (k0xn0) block has to be transposed before been stored.
SimpleTensor< float > src
Copyright (c) 2017-2021 Arm Limited.
void add_option(std::string option)
Adds option to the existing build option list.
const ITensor * get_const_tensor(int id) const
Get constant tensor of a given id.
unsigned int k0
Number of partial accumulations performed by the matrix multiplication.
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.
bool update_window_and_padding(Window &win, Ts &&... patterns)
Update window and padding size for each of the access patterns.
Status validate_image2d_support_on_rhs(const ITensorInfo &tensor_reshaped_info, const GEMMRHSMatrixInfo &rhs_info)
Utility function to validate the image2d OpenCL object support on the RHS reshaped matrix...
GEMM RHS (Right Hand Side) matrix information.
static Status validate(const ITensorInfo *src, const ITensorInfo *dst, const GEMMRHSMatrixInfo &rhs_info)
Static function to check if given info will lead to a valid configuration.
unsigned int n0
Number of columns processed by the matrix multiplication.
TensorShape compute_rhs_reshaped_shape(const ITensorInfo &a, const GEMMRHSMatrixInfo &rhs_info)
Calculate the Right Hand Side matrix reshaped shape.
bool auto_init_if_empty(ITensorInfo &info, const TensorShape &shape, int num_channels, DataType data_type, QuantizationInfo quantization_info=QuantizationInfo())
Auto initialize the tensor info (shape, number of channels and data type) if the current assignment i...
virtual std::unique_ptr< T > clone() const =0
Provide a clone of the current object of class T.
void add_option_if(bool cond, std::string option)
Adds option if a given condition is true;.
virtual size_t element_size() const =0
Element size in bytes calculated as data_size() * num_channels()
ClGemmReshapeRhsMatrixKernel()
Elementeise CL kernel type.
bool slide_window_slice_3D(Window &slice) const
Slide the passed 3D window slice.
#define ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(k)
ITensor * get_tensor(int id)
Get tensor of a given id from the pac.
#define ARM_COMPUTE_CREATE_ERROR(error_code, msg)
Creates an error with a given message.
static constexpr size_t DimZ
Alias for dimension 2 also known as Z dimension.
#define ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(...)
std::string get_cl_unsigned_type_from_element_size(size_t element_size)
Translates the element size to an unsigned integer data type.
Wrapper to configure the Khronos OpenCL C++ header.
bool interleave
True if the h0 (k0xn0) blocks have to be interleaved in the output row.
#define ARM_COMPUTE_RETURN_ERROR_ON_MSG(cond, msg)
If the condition is true, an error is returned.
#define ARM_COMPUTE_ERROR_ON_NULLPTR(...)
Window first_slice_window_3D() const
First 3D slice of the window.
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)