| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | #include "zhurin_i_gaus_kernel/omp/include/ops_omp.hpp" | ||
| 2 | |||
| 3 | #include <algorithm> | ||
| 4 | #include <utility> | ||
| 5 | #include <vector> | ||
| 6 | |||
| 7 | #include "zhurin_i_gaus_kernel/common/include/common.hpp" | ||
| 8 | |||
| 9 | namespace zhurin_i_gaus_kernel { | ||
| 10 | |||
| 11 | 24 | bool ZhurinIGausKernelOMP::ValidationImpl() { | |
| 12 | const auto &in = GetInput(); | ||
| 13 | 24 | int w = std::get<0>(in); | |
| 14 | 24 | int h = std::get<1>(in); | |
| 15 | 24 | int parts = std::get<2>(in); | |
| 16 | const auto &img = std::get<3>(in); | ||
| 17 | |||
| 18 |
2/4✓ Branch 0 taken 24 times.
✗ Branch 1 not taken.
✗ Branch 2 not taken.
✓ Branch 3 taken 24 times.
|
24 | if (w <= 0 || h <= 0 || parts <= 0 || parts > w) { |
| 19 | return false; | ||
| 20 | } | ||
| 21 | if (std::cmp_not_equal(img.size(), h)) { | ||
| 22 | return false; | ||
| 23 | } | ||
| 24 |
2/2✓ Branch 0 taken 64 times.
✓ Branch 1 taken 24 times.
|
88 | for (int i = 0; i < h; ++i) { |
| 25 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 64 times.
|
64 | if (std::cmp_not_equal(img[i].size(), w)) { |
| 26 | return false; | ||
| 27 | } | ||
| 28 | } | ||
| 29 | return true; | ||
| 30 | } | ||
| 31 | |||
| 32 | 24 | bool ZhurinIGausKernelOMP::PreProcessingImpl() { | |
| 33 | const auto &in = GetInput(); | ||
| 34 | 24 | width_ = std::get<0>(in); | |
| 35 | 24 | height_ = std::get<1>(in); | |
| 36 | 24 | num_parts_ = std::get<2>(in); | |
| 37 | |||
| 38 |
1/2✓ Branch 2 taken 24 times.
✗ Branch 3 not taken.
|
24 | padded_.assign(height_ + 2, std::vector<int>(width_ + 2, 0)); |
| 39 | const auto &img = std::get<3>(in); | ||
| 40 |
2/2✓ Branch 0 taken 64 times.
✓ Branch 1 taken 24 times.
|
88 | for (int i = 0; i < height_; ++i) { |
| 41 | 64 | std::copy(img[i].begin(), img[i].end(), padded_[i + 1].begin() + 1); | |
| 42 | } | ||
| 43 | |||
| 44 |
1/2✓ Branch 2 taken 24 times.
✗ Branch 3 not taken.
|
24 | result_.assign(height_, std::vector<int>(width_, 0)); |
| 45 | 24 | return true; | |
| 46 | } | ||
| 47 | |||
| 48 |
1/2✓ Branch 1 taken 24 times.
✗ Branch 2 not taken.
|
24 | ZhurinIGausKernelOMP::ZhurinIGausKernelOMP(const InType &in) { |
| 49 | SetTypeOfTask(GetStaticTypeOfTask()); | ||
| 50 | GetInput() = in; | ||
| 51 | 24 | GetOutput() = OutType{}; | |
| 52 | 24 | } | |
| 53 | |||
| 54 | 24 | bool ZhurinIGausKernelOMP::RunImpl() { | |
| 55 | 24 | int w = width_; | |
| 56 | 24 | int h = height_; | |
| 57 | 24 | int np = num_parts_; | |
| 58 | 24 | int base_width = w / np; | |
| 59 | 24 | int remainder = w % np; | |
| 60 | |||
| 61 | 24 | auto &local_padded = padded_; | |
| 62 | 24 | auto &local_result = result_; | |
| 63 | |||
| 64 | 24 | #pragma omp parallel for schedule(static) default(none) \ | |
| 65 | shared(w, h, base_width, remainder, np, local_padded, local_result) | ||
| 66 | for (int part = 0; part < np; ++part) { | ||
| 67 | int part_width = base_width + (part < remainder ? 1 : 0); | ||
| 68 | int x_start = (part * base_width) + std::min(part, remainder); | ||
| 69 | int x_end = x_start + part_width; | ||
| 70 | |||
| 71 | for (int i = 1; i <= h; ++i) { | ||
| 72 | for (int j = x_start + 1; j <= x_end; ++j) { | ||
| 73 | int sum = (local_padded[i - 1][j - 1] * kKernel[0][0]) + (local_padded[i - 1][j] * kKernel[0][1]) + | ||
| 74 | (local_padded[i - 1][j + 1] * kKernel[0][2]) + (local_padded[i][j - 1] * kKernel[1][0]) + | ||
| 75 | (local_padded[i][j] * kKernel[1][1]) + (local_padded[i][j + 1] * kKernel[1][2]) + | ||
| 76 | (local_padded[i + 1][j - 1] * kKernel[2][0]) + (local_padded[i + 1][j] * kKernel[2][1]) + | ||
| 77 | (local_padded[i + 1][j + 1] * kKernel[2][2]); | ||
| 78 | local_result[i - 1][j - 1] = sum >> kShift; | ||
| 79 | } | ||
| 80 | } | ||
| 81 | } | ||
| 82 | 24 | return true; | |
| 83 | } | ||
| 84 | |||
| 85 | 24 | bool ZhurinIGausKernelOMP::PostProcessingImpl() { | |
| 86 | 24 | GetOutput() = result_; | |
| 87 | 24 | return true; | |
| 88 | } | ||
| 89 | |||
| 90 | } // namespace zhurin_i_gaus_kernel | ||
| 91 |