summaryrefslogtreecommitdiffstats
path: root/contrib/libs/apache/arrow_next/cpp/src/arrow/util/bit_util.h
blob: 6e1f32bf3575c0a156ab19cad1abb2703b886641 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
#pragma clang system_header
// Licensed to the Apache Software Foundation (ASF) under one
// or more contributor license agreements.  See the NOTICE file
// distributed with this work for additional information
// regarding copyright ownership.  The ASF licenses this file
// to you under the Apache License, Version 2.0 (the
// "License"); you may not use this file except in compliance
// with the License.  You may obtain a copy of the License at
//
//   http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing,
// software distributed under the License is distributed on an
// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, either express or implied.  See the License for the
// specific language governing permissions and limitations
// under the License.

#pragma once

#if defined(_MSC_VER)
#  if defined(_M_AMD64) || defined(_M_X64)
#    include <intrin.h>  // IWYU pragma: keep
#  endif

#  pragma intrinsic(_BitScanReverse)
#  pragma intrinsic(_BitScanForward)
#  define ARROW_POPCOUNT64 __popcnt64
#  define ARROW_POPCOUNT32 __popcnt
#else
#  define ARROW_POPCOUNT64 __builtin_popcountll
#  define ARROW_POPCOUNT32 __builtin_popcount
#endif

#include <cstdint>
#include <type_traits>

#include "contrib/libs/apache/arrow_next/cpp/src/arrow/util/macros.h"
#include "contrib/libs/apache/arrow_next/cpp/src/arrow/util/visibility.h"

namespace arrow20 {
namespace detail {

template <typename Integer>
typename std::make_unsigned<Integer>::type as_unsigned(Integer x) {
  return static_cast<typename std::make_unsigned<Integer>::type>(x);
}

}  // namespace detail

namespace bit_util {

// The number of set bits in a given unsigned byte value, pre-computed
//
// Generated with the following Python code
// output = 'static constexpr uint8_t kBytePopcount[] = {{{0}}};'
// popcounts = [str(bin(i).count('1')) for i in range(0, 256)]
// print(output.format(', '.join(popcounts)))
static constexpr uint8_t kBytePopcount[] = {
    0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4, 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3,
    4, 4, 5, 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, 2, 3, 3, 4, 3, 4, 4, 5, 3, 4,
    4, 5, 4, 5, 5, 6, 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, 2, 3, 3, 4, 3, 4, 4,
    5, 3, 4, 4, 5, 4, 5, 5, 6, 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, 3, 4, 4, 5,
    4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, 1, 2, 2, 3, 2, 3, 3, 4, 2, 3, 3, 4, 3, 4, 4, 5, 2,
    3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5, 5, 6, 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4, 5, 4, 5,
    5, 6, 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, 2, 3, 3, 4, 3, 4, 4, 5, 3, 4, 4,
    5, 4, 5, 5, 6, 3, 4, 4, 5, 4, 5, 5, 6, 4, 5, 5, 6, 5, 6, 6, 7, 3, 4, 4, 5, 4, 5, 5, 6,
    4, 5, 5, 6, 5, 6, 6, 7, 4, 5, 5, 6, 5, 6, 6, 7, 5, 6, 6, 7, 6, 7, 7, 8};

static inline uint64_t PopCount(uint64_t bitmap) { return ARROW_POPCOUNT64(bitmap); }
static inline uint32_t PopCount(uint32_t bitmap) { return ARROW_POPCOUNT32(bitmap); }

//
// Bit-related computations on integer values
//

// Returns the ceil of value/divisor
constexpr int64_t CeilDiv(int64_t value, int64_t divisor) {
  return (value == 0) ? 0 : 1 + (value - 1) / divisor;
}

// Return the number of bytes needed to fit the given number of bits
constexpr int64_t BytesForBits(int64_t bits) {
  // This formula avoids integer overflow on very large `bits`
  return (bits >> 3) + ((bits & 7) != 0);
}

constexpr bool IsPowerOf2(int64_t value) {
  return value > 0 && (value & (value - 1)) == 0;
}

constexpr bool IsPowerOf2(uint64_t value) {
  return value > 0 && (value & (value - 1)) == 0;
}

// Returns the smallest power of two that contains v.  If v is already a
// power of two, it is returned as is.
static inline int64_t NextPower2(int64_t n) {
  // Taken from
  // http://graphics.stanford.edu/~seander/bithacks.html#RoundUpPowerOf2
  n--;
  n |= n >> 1;
  n |= n >> 2;
  n |= n >> 4;
  n |= n >> 8;
  n |= n >> 16;
  n |= n >> 32;
  n++;
  return n;
}

constexpr bool IsMultipleOf64(int64_t n) { return (n & 63) == 0; }

constexpr bool IsMultipleOf8(int64_t n) { return (n & 7) == 0; }

// Returns a mask for the bit_index lower order bits.
// Only valid for bit_index in the range [0, 64).
constexpr uint64_t LeastSignificantBitMask(int64_t bit_index) {
  return (static_cast<uint64_t>(1) << bit_index) - 1;
}

// Returns 'value' rounded up to the nearest multiple of 'factor'
constexpr int64_t RoundUp(int64_t value, int64_t factor) {
  return CeilDiv(value, factor) * factor;
}

// Returns 'value' rounded down to the nearest multiple of 'factor'
constexpr int64_t RoundDown(int64_t value, int64_t factor) {
  return (value / factor) * factor;
}

// Returns 'value' rounded up to the nearest multiple of 'factor' when factor
// is a power of two.
// The result is undefined on overflow, i.e. if `value > 2**64 - factor`,
// since we cannot return the correct result which would be 2**64.
constexpr int64_t RoundUpToPowerOf2(int64_t value, int64_t factor) {
  // DCHECK(value >= 0);
  // DCHECK(IsPowerOf2(factor));
  return (value + (factor - 1)) & ~(factor - 1);
}

constexpr uint64_t RoundUpToPowerOf2(uint64_t value, uint64_t factor) {
  // DCHECK(IsPowerOf2(factor));
  return (value + (factor - 1)) & ~(factor - 1);
}

constexpr int64_t RoundUpToMultipleOf8(int64_t num) { return RoundUpToPowerOf2(num, 8); }

constexpr int64_t RoundUpToMultipleOf64(int64_t num) {
  return RoundUpToPowerOf2(num, 64);
}

// Returns the number of bytes covering a sliced bitmap. Find the length
// rounded to cover full bytes on both extremities.
//
// The following example represents a slice (offset=10, length=9)
//
// 0       8       16     24
// |-------|-------|------|
//           [       ]          (slice)
//         [             ]      (same slice aligned to bytes bounds, length=16)
//
// The covering bytes is the length (in bytes) of this new aligned slice.
constexpr int64_t CoveringBytes(int64_t offset, int64_t length) {
  return (bit_util::RoundUp(length + offset, 8) - bit_util::RoundDown(offset, 8)) / 8;
}

// Returns the 'num_bits' least-significant bits of 'v'.
static inline uint64_t TrailingBits(uint64_t v, int num_bits) {
  if (ARROW_PREDICT_FALSE(num_bits == 0)) return 0;
  if (ARROW_PREDICT_FALSE(num_bits >= 64)) return v;
  int n = 64 - num_bits;
  return (v << n) >> n;
}

/// \brief Count the number of leading zeros in an unsigned integer.
static inline int CountLeadingZeros(uint32_t value) {
#if defined(__clang__) || defined(__GNUC__)
  if (value == 0) return 32;
  return static_cast<int>(__builtin_clz(value));
#elif defined(_MSC_VER)
  unsigned long index;                                               // NOLINT
  if (_BitScanReverse(&index, static_cast<unsigned long>(value))) {  // NOLINT
    return 31 - static_cast<int>(index);
  } else {
    return 32;
  }
#else
  int bitpos = 0;
  while (value != 0) {
    value >>= 1;
    ++bitpos;
  }
  return 32 - bitpos;
#endif
}

static inline int CountLeadingZeros(uint64_t value) {
#if defined(__clang__) || defined(__GNUC__)
  if (value == 0) return 64;
  return static_cast<int>(__builtin_clzll(value));
#elif defined(_MSC_VER)
  unsigned long index;                     // NOLINT
  if (_BitScanReverse64(&index, value)) {  // NOLINT
    return 63 - static_cast<int>(index);
  } else {
    return 64;
  }
#else
  int bitpos = 0;
  while (value != 0) {
    value >>= 1;
    ++bitpos;
  }
  return 64 - bitpos;
#endif
}

static inline int CountTrailingZeros(uint32_t value) {
#if defined(__clang__) || defined(__GNUC__)
  if (value == 0) return 32;
  return static_cast<int>(__builtin_ctzl(value));
#elif defined(_MSC_VER)
  unsigned long index;  // NOLINT
  if (_BitScanForward(&index, value)) {
    return static_cast<int>(index);
  } else {
    return 32;
  }
#else
  int bitpos = 0;
  if (value) {
    while (value & 1 == 0) {
      value >>= 1;
      ++bitpos;
    }
  } else {
    bitpos = 32;
  }
  return bitpos;
#endif
}

static inline int CountTrailingZeros(uint64_t value) {
#if defined(__clang__) || defined(__GNUC__)
  if (value == 0) return 64;
  return static_cast<int>(__builtin_ctzll(value));
#elif defined(_MSC_VER)
  unsigned long index;  // NOLINT
  if (_BitScanForward64(&index, value)) {
    return static_cast<int>(index);
  } else {
    return 64;
  }
#else
  int bitpos = 0;
  if (value) {
    while (value & 1 == 0) {
      value >>= 1;
      ++bitpos;
    }
  } else {
    bitpos = 64;
  }
  return bitpos;
#endif
}

// Returns the minimum number of bits needed to represent an unsigned value
static inline int NumRequiredBits(uint64_t x) { return 64 - CountLeadingZeros(x); }

// Returns ceil(log2(x)).
static inline int Log2(uint64_t x) {
  // DCHECK_GT(x, 0);
  return NumRequiredBits(x - 1);
}

//
// Utilities for reading and writing individual bits by their index
// in a memory area.
//

// Bitmask selecting the k-th bit in a byte
static constexpr uint8_t kBitmask[] = {1, 2, 4, 8, 16, 32, 64, 128};

// the bitwise complement version of kBitmask
static constexpr uint8_t kFlippedBitmask[] = {254, 253, 251, 247, 239, 223, 191, 127};

// Bitmask selecting the (k - 1) preceding bits in a byte
static constexpr uint8_t kPrecedingBitmask[] = {0, 1, 3, 7, 15, 31, 63, 127};
static constexpr uint8_t kPrecedingWrappingBitmask[] = {255, 1, 3, 7, 15, 31, 63, 127};

// the bitwise complement version of kPrecedingBitmask
static constexpr uint8_t kTrailingBitmask[] = {255, 254, 252, 248, 240, 224, 192, 128};

static constexpr bool GetBit(const uint8_t* bits, uint64_t i) {
  return (bits[i >> 3] >> (i & 0x07)) & 1;
}

// Gets the i-th bit from a byte. Should only be used with i <= 7.
static constexpr bool GetBitFromByte(uint8_t byte, uint8_t i) {
  return byte & kBitmask[i];
}

static inline void ClearBit(uint8_t* bits, int64_t i) {
  bits[i / 8] &= kFlippedBitmask[i % 8];
}

static inline void SetBit(uint8_t* bits, int64_t i) { bits[i / 8] |= kBitmask[i % 8]; }

static inline void SetBitTo(uint8_t* bits, int64_t i, bool bit_is_set) {
  // https://graphics.stanford.edu/~seander/bithacks.html
  // "Conditionally set or clear bits without branching"
  // NOTE: this seems to confuse Valgrind as it reads from potentially
  // uninitialized memory
  bits[i / 8] ^= static_cast<uint8_t>(-static_cast<uint8_t>(bit_is_set) ^ bits[i / 8]) &
                 kBitmask[i % 8];
}

/// \brief set or clear a range of bits quickly
ARROW_EXPORT
void SetBitsTo(uint8_t* bits, int64_t start_offset, int64_t length, bool bits_are_set);

/// \brief Sets all bits in the bitmap to true
ARROW_EXPORT
void SetBitmap(uint8_t* data, int64_t offset, int64_t length);

/// \brief Clears all bits in the bitmap (set to false)
ARROW_EXPORT
void ClearBitmap(uint8_t* data, int64_t offset, int64_t length);

/// Returns a mask with lower i bits set to 1. If i >= sizeof(Word)*8, all-ones will be
/// returned
/// ex:
/// ref: https://stackoverflow.com/a/59523400
template <typename Word>
constexpr Word PrecedingWordBitmask(unsigned int const i) {
  return static_cast<Word>(static_cast<Word>(i < sizeof(Word) * 8)
                           << (i & (sizeof(Word) * 8 - 1))) -
         1;
}
static_assert(PrecedingWordBitmask<uint8_t>(0) == 0x00, "");
static_assert(PrecedingWordBitmask<uint8_t>(4) == 0x0f, "");
static_assert(PrecedingWordBitmask<uint8_t>(8) == 0xff, "");
static_assert(PrecedingWordBitmask<uint16_t>(8) == 0x00ff, "");

/// \brief Create a word with low `n` bits from `low` and high `sizeof(Word)-n` bits
/// from `high`.
/// Word ret
/// for (i = 0; i < sizeof(Word)*8; i++){
///     ret[i]= i < n ? low[i]: high[i];
/// }
template <typename Word>
constexpr Word SpliceWord(int n, Word low, Word high) {
  return (high & ~PrecedingWordBitmask<Word>(n)) | (low & PrecedingWordBitmask<Word>(n));
}

/// \brief Pack integers into a bitmap in batches of 8
template <int batch_size>
void PackBits(const uint32_t* values, uint8_t* out) {
  for (int i = 0; i < batch_size / 8; ++i) {
    *out++ = static_cast<uint8_t>(values[0] | values[1] << 1 | values[2] << 2 |
                                  values[3] << 3 | values[4] << 4 | values[5] << 5 |
                                  values[6] << 6 | values[7] << 7);
    values += 8;
  }
}

}  // namespace bit_util
}  // namespace arrow20