Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 1 | /* |
Georgios Pinitas | de5a1cc | 2018-02-02 12:52:07 +0000 | [diff] [blame] | 2 | * Copyright (c) 2017-2018 ARM Limited. |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 3 | * |
| 4 | * SPDX-License-Identifier: MIT |
| 5 | * |
| 6 | * Permission is hereby granted, free of charge, to any person obtaining a copy |
| 7 | * of this software and associated documentation files (the "Software"), to |
| 8 | * deal in the Software without restriction, including without limitation the |
| 9 | * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or |
| 10 | * sell copies of the Software, and to permit persons to whom the Software is |
| 11 | * furnished to do so, subject to the following conditions: |
| 12 | * |
| 13 | * The above copyright notice and this permission notice shall be included in all |
| 14 | * copies or substantial portions of the Software. |
| 15 | * |
| 16 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 17 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 18 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
| 19 | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| 20 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, |
| 21 | * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE |
| 22 | * SOFTWARE. |
| 23 | */ |
| 24 | #include "arm_compute/core/CL/kernels/CLDepthwiseVectorToTensorKernel.h" |
| 25 | |
| 26 | #include "arm_compute/core/CL/CLHelpers.h" |
| 27 | #include "arm_compute/core/CL/CLKernelLibrary.h" |
Vidhya Sudhan Loganathan | f1f4906 | 2018-05-25 13:21:26 +0100 | [diff] [blame] | 28 | #include "arm_compute/core/CL/CLValidate.h" |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 29 | #include "arm_compute/core/CL/ICLTensor.h" |
| 30 | #include "arm_compute/core/CL/OpenCL.h" |
| 31 | #include "arm_compute/core/Error.h" |
| 32 | #include "arm_compute/core/Helpers.h" |
| 33 | #include "arm_compute/core/Types.h" |
Abe Mbise | 7784c83 | 2018-05-31 16:48:41 +0100 | [diff] [blame] | 34 | #include "arm_compute/core/utils/misc/ShapeCalculator.h" |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 35 | #include "support/ToolchainSupport.h" |
| 36 | |
| 37 | using namespace arm_compute; |
Abe Mbise | 7784c83 | 2018-05-31 16:48:41 +0100 | [diff] [blame] | 38 | using namespace arm_compute::misc::shape_calculator; |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 39 | |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 40 | namespace |
| 41 | { |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 42 | Status validate_arguments(const ITensorInfo *input, const ITensorInfo *output, size_t conv_w, size_t conv_h) |
| 43 | { |
Vidhya Sudhan Loganathan | f1f4906 | 2018-05-25 13:21:26 +0100 | [diff] [blame] | 44 | ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(input); |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 45 | ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::QASYMM8, DataType::S32, DataType::F16, DataType::F32); |
| 46 | |
| 47 | if(output->total_size() != 0) |
| 48 | { |
Abe Mbise | 7784c83 | 2018-05-31 16:48:41 +0100 | [diff] [blame] | 49 | TensorShape output_shape = compute_vector_to_tensor_output_shape(input->tensor_shape(), conv_w, conv_h, output->data_layout()); |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 50 | ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DIMENSIONS(output->tensor_shape(), output_shape); |
| 51 | ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output); |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 52 | } |
| 53 | |
| 54 | return Status{}; |
| 55 | } |
| 56 | } // namespace |
| 57 | |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 58 | CLDepthwiseVectorToTensorKernel::CLDepthwiseVectorToTensorKernel() |
| 59 | : _input(nullptr), _output(nullptr) |
| 60 | { |
| 61 | } |
| 62 | |
| 63 | void CLDepthwiseVectorToTensorKernel::configure(const ICLTensor *input, ICLTensor *output, size_t conv_w, size_t conv_h) |
| 64 | { |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 65 | ARM_COMPUTE_ERROR_ON_NULLPTR(input, output); |
Georgios Pinitas | 5cbd20f | 2017-10-27 12:01:23 +0100 | [diff] [blame] | 66 | |
| 67 | // Output auto inizialitation if not yet initialized |
Abe Mbise | 7784c83 | 2018-05-31 16:48:41 +0100 | [diff] [blame] | 68 | TensorShape output_shape = compute_vector_to_tensor_output_shape(input->info()->tensor_shape(), conv_w, conv_h, output->info()->data_layout()); |
Georgios Pinitas | 9be0c5a | 2018-02-19 12:46:29 +0000 | [diff] [blame] | 69 | auto_init_if_empty(*output->info(), input->info()->clone()->set_tensor_shape(output_shape)); |
Georgios Pinitas | 5cbd20f | 2017-10-27 12:01:23 +0100 | [diff] [blame] | 70 | |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 71 | ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), output->info(), conv_w, conv_h)); |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 72 | |
| 73 | _input = input; |
| 74 | _output = output; |
| 75 | |
| 76 | // Create kernel |
Georgios Pinitas | de5a1cc | 2018-02-02 12:52:07 +0000 | [diff] [blame] | 77 | CLBuildOptions build_opts; |
| 78 | build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type())); |
| 79 | build_opts.add_option("-DCONV_WIDTH=" + support::cpp11::to_string(conv_w)); |
| 80 | build_opts.add_option("-DCONV_HEIGHT=" + support::cpp11::to_string(conv_h)); |
Giorgio Arena | d051e97 | 2018-06-20 11:46:42 +0100 | [diff] [blame] | 81 | build_opts.add_option("-D" + string_from_data_layout(output->info()->data_layout())); |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 82 | |
Georgios Pinitas | de5a1cc | 2018-02-02 12:52:07 +0000 | [diff] [blame] | 83 | _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("depthwise_vector_to_tensor", build_opts.options())); |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 84 | |
| 85 | // Configure kernel window |
| 86 | Window win = calculate_max_window(*input->info(), Steps()); |
| 87 | // The CLDepthwisevectorToTensorKernel doesn't need padding so update_window_and_padding() can be skipped |
| 88 | output->info()->set_valid_region(ValidRegion(Coordinates(), output->info()->tensor_shape())); |
| 89 | |
Anthony Barbier | b6eb353 | 2018-08-08 13:20:04 +0100 | [diff] [blame] | 90 | ICLKernel::configure_internal(win); |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 91 | } |
| 92 | |
Giorgio Arena | ad0c738 | 2018-04-23 16:16:21 +0100 | [diff] [blame] | 93 | Status CLDepthwiseVectorToTensorKernel::validate(const ITensorInfo *input, const ITensorInfo *output, size_t conv_w, size_t conv_h) |
| 94 | { |
| 95 | ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, output, conv_w, conv_h)); |
| 96 | return Status{}; |
| 97 | } |
| 98 | |
Giorgio Arena | 9fe4144 | 2017-08-23 16:36:24 +0100 | [diff] [blame] | 99 | void CLDepthwiseVectorToTensorKernel::run(const Window &window, cl::CommandQueue &queue) |
| 100 | { |
| 101 | ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this); |
| 102 | ARM_COMPUTE_ERROR_ON_MISMATCHING_WINDOWS(ICLKernel::window(), window); |
| 103 | |
| 104 | Window slice = window.first_slice_window_1D(); |
| 105 | Window slice_out = window.first_slice_window_3D(); |
| 106 | |
| 107 | // Setup slice |
| 108 | slice.set(Window::DimX, Window::Dimension(0, _input->info()->dimension(0), 1)); |
| 109 | |
| 110 | // Setup output slice |
| 111 | // The first three dimensions of the output are increased by the inner loops |
| 112 | slice_out.set(Window::DimX, Window::Dimension(0, 0, 0)); |
| 113 | slice_out.set(Window::DimY, Window::Dimension(0, 0, 0)); |
| 114 | slice_out.set(Window::DimZ, Window::Dimension(0, 0, 0)); |
| 115 | |
| 116 | do |
| 117 | { |
| 118 | unsigned int idx = 0; |
| 119 | add_1D_tensor_argument(idx, _input, slice); |
| 120 | add_3D_tensor_argument(idx, _output, slice_out); |
| 121 | enqueue(queue, *this, slice); |
| 122 | } |
| 123 | while(window.slide_window_slice_1D(slice) && window.slide_window_slice_3D(slice_out)); |
| 124 | } |