diff options
author | steniu01 <steven.niu@arm.com> | 2017-07-17 23:16:00 +0100 |
---|---|---|
committer | Anthony Barbier <anthony.barbier@arm.com> | 2018-11-02 16:35:24 +0000 |
commit | 868e54102b1761577e7783c80b66e82896ee063c (patch) | |
tree | 8578c4fc3b44c867ae9c8529a3b21730a5ce14f3 /src/core/CL | |
parent | 5cb4d6a1d0f39bf800edb43c0ec7c96dae10e132 (diff) | |
download | ComputeLibrary-868e54102b1761577e7783c80b66e82896ee063c.tar.gz |
COMPMID-459 Collapse CL Im2col's higher dimensions
Change-Id: I0ccc39cbcf6926e6810faf3fe264c4af7adc3f7b
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/83070
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
Diffstat (limited to 'src/core/CL')
-rw-r--r-- | src/core/CL/ICLKernel.cpp | 10 | ||||
-rw-r--r-- | src/core/CL/cl_kernels/convolution_layer.cl | 37 | ||||
-rw-r--r-- | src/core/CL/cl_kernels/helpers.h | 64 | ||||
-rw-r--r-- | src/core/CL/kernels/CLIm2ColKernel.cpp | 24 |
4 files changed, 114 insertions, 21 deletions
diff --git a/src/core/CL/ICLKernel.cpp b/src/core/CL/ICLKernel.cpp index 5bd7142176..bace631549 100644 --- a/src/core/CL/ICLKernel.cpp +++ b/src/core/CL/ICLKernel.cpp @@ -122,6 +122,11 @@ void ICLKernel::add_3D_tensor_argument(unsigned int &idx, const ICLTensor *tenso add_tensor_argument<3>(idx, tensor, window); } +void ICLKernel::add_4D_tensor_argument(unsigned int &idx, const ICLTensor *tensor, const Window &window) +{ + add_tensor_argument<4>(idx, tensor, window); +} + unsigned int ICLKernel::num_arguments_per_1D_tensor() const { return num_arguments_per_tensor<1>(); @@ -137,6 +142,11 @@ unsigned int ICLKernel::num_arguments_per_3D_tensor() const return num_arguments_per_tensor<3>(); } +unsigned int ICLKernel::num_arguments_per_4D_tensor() const +{ + return num_arguments_per_tensor<4>(); +} + void ICLKernel::set_target(cl::Device &device) { _target = get_target_from_device(device); diff --git a/src/core/CL/cl_kernels/convolution_layer.cl b/src/core/CL/cl_kernels/convolution_layer.cl index 0dd331f171..cd886b1b0f 100644 --- a/src/core/CL/cl_kernels/convolution_layer.cl +++ b/src/core/CL/cl_kernels/convolution_layer.cl @@ -120,11 +120,15 @@ __kernel void reshape_to_columns( */ __kernel void im2col_generic( TENSOR3D_DECLARATION(src), - IMAGE_DECLARATION(dst)) + IMAGE_DECLARATION(dst), + uint filter_depth, + uint src_stride_w, + uint dst_stride_w) { - const int xc = get_global_id(0); // x coordinate in the convolved tensor - const int yc = get_global_id(1); // y coordinate in the convolved tensor - const int ch = get_global_id(2); // input feature map + const int xc = get_global_id(0); // x coordinate in the convolved tensor + const int yc = get_global_id(1); // y coordinate in the convolved tensor + const int ch = get_global_id(2) % filter_depth; // input feature map + const int batch = get_global_id(2) / filter_depth; // the batch // Calculate input indeces const int xi = xc * STRIDE_X - PAD_X; @@ -134,8 +138,8 @@ __kernel void im2col_generic( const int xo = ch * KERNEL_WIDTH * KERNEL_HEIGHT; const int yo = xc + yc * CONVOLVED_WIDTH; // Index of the convolution - __global uchar *input_ptr = src_ptr + src_offset_first_element_in_bytes + ch * src_stride_z; - __global DATA_TYPE *output_ptr = ((__global DATA_TYPE *)(dst_ptr + dst_offset_first_element_in_bytes + yo * dst_stride_y)) + xo; + __global uchar *input_ptr = src_ptr + src_offset_first_element_in_bytes + ch * src_stride_z + batch * src_stride_w; + __global DATA_TYPE *output_ptr = ((__global DATA_TYPE *)(dst_ptr + dst_offset_first_element_in_bytes + yo * dst_stride_y + batch * dst_stride_w)) + xo; // Linearize convolution elements for(int y = yi, y_e = yi + KERNEL_HEIGHT; y < y_e; ++y) @@ -158,7 +162,7 @@ __kernel void im2col_generic( } #ifdef HAS_BIAS - if(get_global_id(2) == (KERNEL_DEPTH - 1)) + if(ch == (KERNEL_DEPTH - 1)) { #ifdef FIXED_POINT_POSITION *output_ptr = (DATA_TYPE)(1 << FIXED_POINT_POSITION); @@ -191,11 +195,15 @@ __kernel void im2col_generic( */ __kernel void im2col_kernel3x3_padx0_pady0( TENSOR3D_DECLARATION(src), - IMAGE_DECLARATION(dst)) + IMAGE_DECLARATION(dst), + uint filter_depth, + uint src_stride_w, + uint dst_stride_w) { - const int xc = get_global_id(0); // x coordinate in the convolved tensor - const int yc = get_global_id(1); // y coordinate in the convolved tensor - const int ch = get_global_id(2); // input feature map + const int xc = get_global_id(0); // x coordinate in the convolved tensor + const int yc = get_global_id(1); // y coordinate in the convolved tensor + const int ch = get_global_id(2) % filter_depth; // input feature map + const int batch = get_global_id(2) / filter_depth; // the batch // Calculate input indeces const int xi = xc * STRIDE_X; @@ -206,8 +214,9 @@ __kernel void im2col_kernel3x3_padx0_pady0( const int yo = xc + yc * CONVOLVED_WIDTH; // Index of the convolution // Get input and output address - __global uchar *input_ptr = src_ptr + src_offset_first_element_in_bytes + xi * src_stride_x + yi * src_stride_y + ch * src_stride_z; - __global DATA_TYPE *output_ptr = ((__global DATA_TYPE *)(dst_ptr + dst_offset_first_element_in_bytes + yo * dst_stride_y)) + xo; + __global uchar *input_ptr = src_ptr + src_offset_first_element_in_bytes + xi * src_stride_x + yi * src_stride_y + ch * src_stride_z + batch * src_stride_w; + + __global DATA_TYPE *output_ptr = (__global DATA_TYPE *)(dst_ptr + dst_offset_first_element_in_bytes + yo * dst_stride_y + batch * dst_stride_w) + xo; VEC_DATA_TYPE(DATA_TYPE, 3) row0 = vload3(0, (__global DATA_TYPE *)(input_ptr + 0 * src_stride_y)); @@ -220,7 +229,7 @@ __kernel void im2col_kernel3x3_padx0_pady0( *(output_ptr + 8) = row2.s2; #ifdef HAS_BIAS - if(get_global_id(2) == (KERNEL_DEPTH - 1)) + if(ch == (KERNEL_DEPTH - 1)) { #ifdef FIXED_POINT_POSITION *(output_ptr + 9) = (DATA_TYPE)(1 << FIXED_POINT_POSITION); diff --git a/src/core/CL/cl_kernels/helpers.h b/src/core/CL/cl_kernels/helpers.h index 59b81d7f06..68af64e344 100644 --- a/src/core/CL/cl_kernels/helpers.h +++ b/src/core/CL/cl_kernels/helpers.h @@ -72,6 +72,18 @@ uint name##_step_z, \ uint name##_offset_first_element_in_bytes +#define TENSOR4D_DECLARATION(name) \ + __global uchar *name##_ptr, \ + uint name##_stride_x, \ + uint name##_step_x, \ + uint name##_stride_y, \ + uint name##_step_y, \ + uint name##_stride_z, \ + uint name##_step_z, \ + uint name##_stride_w, \ + uint name##_step_w, \ + uint name##_offset_first_element_in_bytes + #define CONVERT_TO_VECTOR_STRUCT(name) \ update_vector_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x) @@ -84,6 +96,9 @@ #define CONVERT_TO_IMAGE_STRUCT_NO_STEP(name) \ update_image_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0) +#define CONVERT_TENSOR3D_TO_IMAGE_STRUCT(name) \ + update_image_from_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y, name##_stride_z, name##_step_z) + #define CONVERT_TENSOR3D_TO_IMAGE_STRUCT_NO_STEP(name) \ update_image_from_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, name##_step_z) @@ -97,6 +112,13 @@ #define CONVERT_TO_TENSOR3D_STRUCT_NO_STEP(name) \ update_tensor3D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0) +#define CONVERT_TO_TENSOR4D_STRUCT(name, mod_size) \ + update_tensor4D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y, \ + name##_stride_z, name##_step_z, name##_stride_w, name##_step_z, mod_size) + +#define CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(name, mod_size) \ + update_tensor4D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0, name##_stride_w, 0, mod_size) + /** Structure to hold Vector information */ typedef struct Vector { @@ -124,6 +146,17 @@ typedef struct Tensor3D int stride_z; /**< Stride of the image in Z dimension (in bytes) */ } Tensor3D; +/** Structure to hold 4D tensor information */ +typedef struct Tensor4D +{ + __global uchar *ptr; /**< Pointer to the starting postion of the buffer */ + int offset_first_element_in_bytes; /**< The offset of the first element in the source image */ + int stride_x; /**< Stride of the image in X dimension (in bytes) */ + int stride_y; /**< Stride of the image in Y dimension (in bytes) */ + int stride_z; /**< Stride of the image in Z dimension (in bytes) */ + int stride_w; /**< Stride of the image in W dimension (in bytes) */ +} Tensor4D; + /** Wrap vector information into an Vector structure, and make the pointer point at this workitem's data. * * @param[in] ptr Pointer to the starting postion of the buffer @@ -222,6 +255,24 @@ Tensor3D inline update_tensor3D_workitem_ptr(__global uchar *ptr, uint offset_fi return tensor; } +Tensor4D inline update_tensor4D_workitem_ptr(__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z, uint stride_w, + uint step_w, + uint mod_size) +{ + Tensor4D tensor = + { + .ptr = ptr, + .offset_first_element_in_bytes = offset_first_element_in_bytes, + .stride_x = stride_x, + .stride_y = stride_y, + .stride_z = stride_z, + .stride_w = stride_w + }; + + tensor.ptr += tensor.offset_first_element_in_bytes + get_global_id(0) * step_x + get_global_id(1) * step_y + (get_global_id(2) % mod_size) * step_z + (get_global_id(2) / mod_size) * step_w; + return tensor; +} + /** Get the pointer position of a Vector * * @param[in] vec Pointer to the starting position of the buffer @@ -255,4 +306,17 @@ __global inline const uchar *tensor3D_offset(const Tensor3D *tensor, int x, int return tensor->ptr + x * tensor->stride_x + y * tensor->stride_y + z * tensor->stride_z; } +/** Get the pointer position of a Tensor4D + * + * @param[in] tensor Pointer to the starting position of the buffer + * @param[in] x Relative X position + * @param[in] y Relative Y position + * @param[in] z Relative Z position + * @param[in] w Relative W position + */ +__global inline const uchar *tensor4D_offset(const Tensor4D *tensor, int x, int y, int z, int w) +{ + return tensor->ptr + x * tensor->stride_x + y * tensor->stride_y + z * tensor->stride_z + w * tensor->stride_w; +} + #endif // _HELPER_H diff --git a/src/core/CL/kernels/CLIm2ColKernel.cpp b/src/core/CL/kernels/CLIm2ColKernel.cpp index b72aff26c6..5147ea0609 100644 --- a/src/core/CL/kernels/CLIm2ColKernel.cpp +++ b/src/core/CL/kernels/CLIm2ColKernel.cpp @@ -103,7 +103,6 @@ void CLIm2ColKernel::configure(const ICLTensor *input, ICLTensor *output, const { _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("im2col_generic", build_opts)); } - _run_func = &CLIm2ColKernel::run_generic; } else @@ -117,6 +116,12 @@ void CLIm2ColKernel::configure(const ICLTensor *input, ICLTensor *output, const Window win = calculate_max_window(*input->info(), Steps()); // The CLIm2ColKernel doesn't need padding so update_window_and_padding() can be skipped output->info()->set_valid_region(ValidRegion(Coordinates(), output->info()->tensor_shape())); + if(!run_img2col_reduced) + { + // set the Z dimension's step same size as the whole dimension so that one can't split across the Z dimension + win.set_dimension_step(Window::DimZ, win[Window::DimZ].end() - win[Window::DimZ].start()); + } + ICLKernel::configure(win); } @@ -132,14 +137,17 @@ void CLIm2ColKernel::run_generic(const Window &window, cl::CommandQueue &queue) ARM_COMPUTE_ERROR_ON_MISMATCHING_WINDOWS(ICLKernel::window(), window); // Get initial windows - Window slice = window.first_slice_window_3D(); - Window slice_in = window.first_slice_window_3D(); - Window slice_out = window.first_slice_window_3D(); + Window window_collapsed = window.collapse_if_possible(ICLKernel::window(), Window::DimZ); + // Change the Z dimension's step back to 1 + window_collapsed.set_dimension_step(Window::DimZ, 1); + + Window slice = window_collapsed.first_slice_window_3D(); + Window slice_in = window_collapsed.first_slice_window_3D(); + Window slice_out = window_collapsed.first_slice_window_3D(); // Setup slice slice.set(Window::DimX, Window::Dimension(0, static_cast<int>(_convolved_dims.first), 1)); slice.set(Window::DimY, Window::Dimension(0, static_cast<int>(_convolved_dims.second), 1)); - slice.set(Window::DimZ, Window::Dimension(0, static_cast<int>(_input->info()->dimension(2)), 1)); // Setup input slice // The first three dimensions of the input are increased by the inner loops @@ -157,13 +165,15 @@ void CLIm2ColKernel::run_generic(const Window &window, cl::CommandQueue &queue) do { - // Set inputs unsigned int idx = 0; add_3D_tensor_argument(idx, _input, slice_in); add_2D_tensor_argument(idx, _output, slice_out); + _kernel.setArg<cl_uint>(idx++, static_cast<unsigned int>(_input->info()->dimension(2))); + _kernel.setArg<cl_uint>(idx++, static_cast<unsigned int>(_input->info()->strides_in_bytes()[3])); + _kernel.setArg<cl_uint>(idx++, static_cast<unsigned int>(_output->info()->strides_in_bytes()[3])); enqueue(queue, *this, slice, _lws_hint); } - while(window.slide_window_slice_3D(slice) && window.slide_window_slice_3D(slice_out) && window.slide_window_slice_3D(slice_in)); + while(window_collapsed.slide_window_slice_3D(slice) && window_collapsed.slide_window_slice_3D(slice_out) && window_collapsed.slide_window_slice_3D(slice_in)); } void CLIm2ColKernel::run_reduced(const Window &window, cl::CommandQueue &queue) |