| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | #include "akimov_i_radixsort_int_merge/all/include/ops_all.hpp" | ||
| 2 | |||
| 3 | #include <mpi.h> | ||
| 4 | #include <oneapi/tbb/parallel_for.h> | ||
| 5 | |||
| 6 | #include <algorithm> | ||
| 7 | #include <cstddef> | ||
| 8 | #include <cstdint> | ||
| 9 | #include <iterator> | ||
| 10 | #include <vector> | ||
| 11 | |||
| 12 | #include "akimov_i_radixsort_int_merge/common/include/common.hpp" | ||
| 13 | #include "util/include/util.hpp" | ||
| 14 | |||
| 15 | namespace akimov_i_radixsort_int_merge { | ||
| 16 | |||
| 17 | namespace { | ||
| 18 | |||
| 19 | 24 | void CountingSortStep(std::vector<int>::iterator in_begin, std::vector<int>::iterator in_end, | |
| 20 | std::vector<int>::iterator out_begin, size_t byte_index) { | ||
| 21 | 24 | size_t shift = byte_index * 8; | |
| 22 | 24 | std::vector<size_t> count(256, 0); | |
| 23 | |||
| 24 |
2/2✓ Branch 0 taken 4440 times.
✓ Branch 1 taken 24 times.
|
4464 | for (auto it = in_begin; it != in_end; ++it) { |
| 25 | 4440 | auto raw_val = static_cast<uint32_t>(*it); | |
| 26 | 4440 | uint32_t byte_val = (raw_val >> shift) & 0xFF; | |
| 27 | 4440 | ++count[byte_val]; | |
| 28 | } | ||
| 29 | |||
| 30 |
1/4✓ Branch 1 taken 24 times.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
✗ Branch 4 not taken.
|
24 | std::vector<size_t> prefix(256, 0); |
| 31 | 24 | prefix[0] = 0; | |
| 32 |
2/2✓ Branch 0 taken 6120 times.
✓ Branch 1 taken 24 times.
|
6144 | for (int i = 1; i < 256; ++i) { |
| 33 | 6120 | prefix[i] = prefix[i - 1] + count[i - 1]; | |
| 34 | } | ||
| 35 | |||
| 36 |
2/2✓ Branch 0 taken 4440 times.
✓ Branch 1 taken 24 times.
|
4464 | for (auto it = in_begin; it != in_end; ++it) { |
| 37 | 4440 | auto raw_val = static_cast<uint32_t>(*it); | |
| 38 | 4440 | uint32_t byte_val = (raw_val >> shift) & 0xFF; | |
| 39 | 4440 | *(out_begin + static_cast<std::ptrdiff_t>(prefix[byte_val])) = *it; | |
| 40 | 4440 | ++prefix[byte_val]; | |
| 41 | } | ||
| 42 | 24 | } | |
| 43 | |||
| 44 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 6 times.
|
6 | void RadixSortLocal(std::vector<int>::iterator begin, std::vector<int>::iterator end) { |
| 45 | 6 | size_t n = std::distance(begin, end); | |
| 46 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 6 times.
|
6 | if (n < 2) { |
| 47 | ✗ | return; | |
| 48 | } | ||
| 49 | 6 | std::vector<int> temp(n); | |
| 50 |
2/2✓ Branch 0 taken 24 times.
✓ Branch 1 taken 6 times.
|
30 | for (size_t i = 0; i < sizeof(int); ++i) { |
| 51 |
2/2✓ Branch 0 taken 12 times.
✓ Branch 1 taken 12 times.
|
24 | if (i % 2 == 0) { |
| 52 |
1/2✓ Branch 1 taken 12 times.
✗ Branch 2 not taken.
|
12 | CountingSortStep(begin, end, temp.begin(), i); |
| 53 | } else { | ||
| 54 |
1/2✓ Branch 1 taken 12 times.
✗ Branch 2 not taken.
|
12 | CountingSortStep(temp.begin(), temp.end(), begin, i); |
| 55 | } | ||
| 56 | } | ||
| 57 | } | ||
| 58 | |||
| 59 | 3 | void ParallelMerge(std::vector<int> &arr, const std::vector<int> &offsets, int num_blocks) { | |
| 60 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | for (int step = 1; step < num_blocks; step *= 2) { |
| 61 | 3 | tbb::parallel_for(0, num_blocks, [&, step](int i) { | |
| 62 |
3/4✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
✓ Branch 2 taken 3 times.
✗ Branch 3 not taken.
|
6 | if (i % (2 * step) == 0 && i + step < num_blocks) { |
| 63 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | auto begin = arr.begin() + offsets[i]; |
| 64 | 3 | auto middle = arr.begin() + offsets[i + step]; | |
| 65 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 3 times.
|
3 | int end_idx = std::min(i + (2 * step), num_blocks); |
| 66 | 3 | auto end = arr.begin() + offsets[end_idx]; | |
| 67 | 3 | std::inplace_merge(begin, middle, end); | |
| 68 | } | ||
| 69 | 6 | }); | |
| 70 | } | ||
| 71 | 3 | } | |
| 72 | |||
| 73 | 6 | void DistributeData(int rank, int world_size, int n, const std::vector<int> &input, std::vector<int> &local_data, | |
| 74 | std::vector<int> &send_counts, std::vector<int> &send_displs) { | ||
| 75 | 6 | send_counts.resize(world_size); | |
| 76 | 6 | send_displs.resize(world_size); | |
| 77 | 6 | int base = n / world_size; | |
| 78 | 6 | int rem = n % world_size; | |
| 79 | int offset = 0; | ||
| 80 |
2/2✓ Branch 0 taken 12 times.
✓ Branch 1 taken 6 times.
|
18 | for (int i = 0; i < world_size; ++i) { |
| 81 |
1/2✓ Branch 0 taken 12 times.
✗ Branch 1 not taken.
|
24 | send_counts[i] = base + (i < rem ? 1 : 0); |
| 82 | 12 | send_displs[i] = offset; | |
| 83 | 12 | offset += send_counts[i]; | |
| 84 | } | ||
| 85 | |||
| 86 | 6 | int local_size = send_counts[rank]; | |
| 87 | 6 | local_data.resize(local_size); | |
| 88 | |||
| 89 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | if (ppc::util::IsUnderMpirun()) { |
| 90 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
9 | MPI_Scatterv(rank == 0 ? input.data() : nullptr, send_counts.data(), send_displs.data(), MPI_INT, local_data.data(), |
| 91 | local_size, MPI_INT, 0, MPI_COMM_WORLD); | ||
| 92 | } else { | ||
| 93 | ✗ | local_data = input; | |
| 94 | } | ||
| 95 | 6 | } | |
| 96 | |||
| 97 | 6 | void GatherData(int rank, int n, const std::vector<int> &local_data, const std::vector<int> &send_counts, | |
| 98 | const std::vector<int> &send_displs, std::vector<int> &output) { | ||
| 99 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | if (ppc::util::IsUnderMpirun()) { |
| 100 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | if (rank == 0) { |
| 101 | 3 | output.resize(n); | |
| 102 | } | ||
| 103 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
9 | MPI_Gatherv(local_data.data(), static_cast<int>(local_data.size()), MPI_INT, rank == 0 ? output.data() : nullptr, |
| 104 | send_counts.data(), send_displs.data(), MPI_INT, 0, MPI_COMM_WORLD); | ||
| 105 | } else { | ||
| 106 | ✗ | output = local_data; | |
| 107 | } | ||
| 108 | 6 | } | |
| 109 | |||
| 110 | void InvertSignBit(std::vector<int> &data, int32_t mask) { | ||
| 111 |
4/4✓ Branch 0 taken 1110 times.
✓ Branch 1 taken 6 times.
✓ Branch 2 taken 1110 times.
✓ Branch 3 taken 6 times.
|
2232 | for (int &val : data) { |
| 112 | 2220 | val ^= mask; | |
| 113 | } | ||
| 114 | } | ||
| 115 | |||
| 116 | } // namespace | ||
| 117 | |||
| 118 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | AkimovIRadixSortIntMergeALL::AkimovIRadixSortIntMergeALL(const InType &in) { |
| 119 | SetTypeOfTask(GetStaticTypeOfTask()); | ||
| 120 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | GetInput() = in; |
| 121 | 6 | GetOutput() = OutType(); | |
| 122 | 6 | } | |
| 123 | |||
| 124 | 6 | bool AkimovIRadixSortIntMergeALL::ValidationImpl() { | |
| 125 | 6 | return !GetInput().empty(); | |
| 126 | } | ||
| 127 | |||
| 128 | 6 | bool AkimovIRadixSortIntMergeALL::PreProcessingImpl() { | |
| 129 | 6 | int rank = 0; | |
| 130 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | if (ppc::util::IsUnderMpirun()) { |
| 131 | 6 | MPI_Comm_rank(MPI_COMM_WORLD, &rank); | |
| 132 | } | ||
| 133 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | if (rank == 0) { |
| 134 | 3 | GetOutput() = GetInput(); | |
| 135 | } | ||
| 136 | 6 | return true; | |
| 137 | } | ||
| 138 | |||
| 139 | 6 | bool AkimovIRadixSortIntMergeALL::RunImpl() { | |
| 140 | 6 | const bool is_mpi = ppc::util::IsUnderMpirun(); | |
| 141 | 6 | int rank = 0; | |
| 142 | 6 | int world_size = 1; | |
| 143 | |||
| 144 |
1/2✓ Branch 0 taken 6 times.
✗ Branch 1 not taken.
|
6 | if (is_mpi) { |
| 145 | 6 | MPI_Comm_rank(MPI_COMM_WORLD, &rank); | |
| 146 | 6 | MPI_Comm_size(MPI_COMM_WORLD, &world_size); | |
| 147 | } | ||
| 148 | |||
| 149 | 6 | int n = 0; | |
| 150 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | if (rank == 0) { |
| 151 | 3 | n = static_cast<int>(GetOutput().size()); | |
| 152 | } | ||
| 153 | |||
| 154 |
1/2✓ Branch 0 taken 6 times.
✗ Branch 1 not taken.
|
6 | if (is_mpi) { |
| 155 | 6 | MPI_Bcast(&n, 1, MPI_INT, 0, MPI_COMM_WORLD); | |
| 156 | } | ||
| 157 | |||
| 158 |
1/2✓ Branch 0 taken 6 times.
✗ Branch 1 not taken.
|
6 | if (n == 0) { |
| 159 | return true; | ||
| 160 | } | ||
| 161 | |||
| 162 | 6 | std::vector<int> send_counts; | |
| 163 | 6 | std::vector<int> send_displs; | |
| 164 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | std::vector<int> local_data; |
| 165 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | DistributeData(rank, world_size, n, GetOutput(), local_data, send_counts, send_displs); |
| 166 | |||
| 167 | constexpr int32_t kSignMask = INT32_MIN; | ||
| 168 | InvertSignBit(local_data, kSignMask); | ||
| 169 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | RadixSortLocal(local_data.begin(), local_data.end()); |
| 170 | InvertSignBit(local_data, kSignMask); | ||
| 171 | |||
| 172 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | GatherData(rank, n, local_data, send_counts, send_displs, GetOutput()); |
| 173 | |||
| 174 |
3/4✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
✓ Branch 2 taken 3 times.
✗ Branch 3 not taken.
|
6 | if (rank == 0 && world_size > 1) { |
| 175 |
1/4✓ Branch 1 taken 3 times.
✗ Branch 2 not taken.
✗ Branch 3 not taken.
✗ Branch 4 not taken.
|
3 | std::vector<int> offsets(world_size + 1, 0); |
| 176 |
2/2✓ Branch 0 taken 6 times.
✓ Branch 1 taken 3 times.
|
9 | for (int i = 0; i < world_size; ++i) { |
| 177 | 6 | offsets[i + 1] = offsets[i] + send_counts[i]; | |
| 178 | } | ||
| 179 |
1/2✓ Branch 1 taken 3 times.
✗ Branch 2 not taken.
|
3 | ParallelMerge(GetOutput(), offsets, world_size); |
| 180 | } | ||
| 181 | |||
| 182 |
1/2✓ Branch 0 taken 6 times.
✗ Branch 1 not taken.
|
6 | if (is_mpi) { |
| 183 | 6 | int output_size = static_cast<int>(GetOutput().size()); | |
| 184 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | MPI_Bcast(&output_size, 1, MPI_INT, 0, MPI_COMM_WORLD); |
| 185 |
2/2✓ Branch 0 taken 3 times.
✓ Branch 1 taken 3 times.
|
6 | if (rank != 0) { |
| 186 |
1/2✓ Branch 1 taken 3 times.
✗ Branch 2 not taken.
|
3 | GetOutput().resize(output_size); |
| 187 | } | ||
| 188 |
1/2✓ Branch 1 taken 6 times.
✗ Branch 2 not taken.
|
6 | MPI_Bcast(GetOutput().data(), output_size, MPI_INT, 0, MPI_COMM_WORLD); |
| 189 | } | ||
| 190 | |||
| 191 | return true; | ||
| 192 | } | ||
| 193 | |||
| 194 | 6 | bool AkimovIRadixSortIntMergeALL::PostProcessingImpl() { | |
| 195 | 6 | return true; | |
| 196 | } | ||
| 197 | |||
| 198 | } // namespace akimov_i_radixsort_int_merge | ||
| 199 |