1 /*
2 * Copyright (c) 2018-2020 Arm Limited.
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 "src/core/NEON/kernels/NEMemsetKernel.h"
25
26 #include "arm_compute/core/Helpers.h"
27 #include "arm_compute/core/ITensor.h"
28 #include "arm_compute/core/Utils.h"
29 #include "arm_compute/core/Validate.h"
30 #include "arm_compute/core/Window.h"
31 #include "src/core/AccessWindowStatic.h"
32 #include "src/core/helpers/AutoConfiguration.h"
33 #include "src/core/helpers/WindowHelpers.h"
34
35 namespace arm_compute
36 {
NEMemsetKernel()37 NEMemsetKernel::NEMemsetKernel()
38 : _tensor(nullptr), _constant_value()
39 {
40 }
41
configure(ITensor * tensor,const PixelValue & constant_value)42 void NEMemsetKernel::configure(ITensor *tensor, const PixelValue &constant_value)
43 {
44 ARM_COMPUTE_ERROR_ON_NULLPTR(tensor);
45 _tensor = tensor;
46 _constant_value = constant_value;
47
48 // Configure kernel window
49 Window win = calculate_max_window(*tensor->info(), Steps());
50 INEKernel::configure(win);
51 }
52
run(const Window & window,const ThreadInfo & info)53 void NEMemsetKernel::run(const Window &window, const ThreadInfo &info)
54 {
55 ARM_COMPUTE_UNUSED(info);
56 ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
57 ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(INEKernel::window(), window);
58
59 // Collapse all the batches on the third dimension
60 bool has_collapsed = true;
61 Window collapsed = window.collapse_if_possible(window, Window::DimZ, &has_collapsed);
62 ARM_COMPUTE_ERROR_ON(!has_collapsed);
63
64 uint8_t *const start_valid_region = _tensor->ptr_to_element(_tensor->info()->valid_region().anchor);
65 const auto window_width = static_cast<int>(collapsed.x().end()) - static_cast<int>(collapsed.x().start());
66 const size_t element_size = _tensor->info()->element_size();
67
68 // Unroll X dimension
69 collapsed.set(Window::DimX, Window::Dimension(0, 1, 1));
70
71 Iterator tensor_it(_tensor, collapsed);
72 execute_window_loop(collapsed, [&](const Coordinates &)
73 {
74 uint8_t *base_addr = start_valid_region + tensor_it.offset();
75 // Set memory
76 for(int i = 0; i < window_width; ++i)
77 {
78 std::memcpy(base_addr + i * element_size, &_constant_value.value, element_size);
79 }
80
81 },
82 tensor_it);
83 }
84 } // namespace arm_compute
85