aboutsummaryrefslogtreecommitdiffhomepage
path: root/tensorflow/core/lib/hash/hash.cc
blob: 075d2524120c477bd0c370db5ad653589c63ed35 (plain)
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
#include "tensorflow/core/lib/hash/hash.h"

#include "tensorflow/core/platform/port.h"
#include "tensorflow/core/lib/core/raw_coding.h"

#include <string.h>

namespace tensorflow {

// 0xff is in case char is signed.
static inline uint32 ByteAs32(char c) { return static_cast<uint32>(c) & 0xff; }
static inline uint64 ByteAs64(char c) { return static_cast<uint64>(c) & 0xff; }

uint32 Hash32(const char* data, size_t n, uint32 seed) {
  // 'm' and 'r' are mixing constants generated offline.
  // They're not really 'magic', they just happen to work well.

  const uint32 m = 0x5bd1e995;
  const int r = 24;

  // Initialize the hash to a 'random' value
  uint32 h = seed ^ n;

  // Mix 4 bytes at a time into the hash
  while (n >= 4) {
    uint32 k = core::DecodeFixed32(data);

    k *= m;
    k ^= k >> r;
    k *= m;

    h *= m;
    h ^= k;

    data += 4;
    n -= 4;
  }

  // Handle the last few bytes of the input array

  switch (n) {
    case 3:
      h ^= ByteAs32(data[2]) << 16;
      TF_FALLTHROUGH_INTENDED;
    case 2:
      h ^= ByteAs32(data[1]) << 8;
      TF_FALLTHROUGH_INTENDED;
    case 1:
      h ^= ByteAs32(data[0]);
      h *= m;
  }

  // Do a few final mixes of the hash to ensure the last few
  // bytes are well-incorporated.

  h ^= h >> 13;
  h *= m;
  h ^= h >> 15;

  return h;
}

uint64 Hash64(const char* data, size_t n, uint64 seed) {
  const uint64 m = 0xc6a4a7935bd1e995;
  const int r = 47;

  uint64 h = seed ^ (n * m);

  while (n >= 8) {
    uint64 k = core::DecodeFixed64(data);
    data += 8;
    n -= 8;

    k *= m;
    k ^= k >> r;
    k *= m;

    h ^= k;
    h *= m;
  }

  switch (n) {
    case 7:
      h ^= ByteAs64(data[6]) << 48;
      TF_FALLTHROUGH_INTENDED;
    case 6:
      h ^= ByteAs64(data[5]) << 40;
      TF_FALLTHROUGH_INTENDED;
    case 5:
      h ^= ByteAs64(data[4]) << 32;
      TF_FALLTHROUGH_INTENDED;
    case 4:
      h ^= ByteAs64(data[3]) << 24;
      TF_FALLTHROUGH_INTENDED;
    case 3:
      h ^= ByteAs64(data[2]) << 16;
      TF_FALLTHROUGH_INTENDED;
    case 2:
      h ^= ByteAs64(data[1]) << 8;
      TF_FALLTHROUGH_INTENDED;
    case 1:
      h ^= ByteAs64(data[0]);
      h *= m;
  }

  h ^= h >> r;
  h *= m;
  h ^= h >> r;

  return h;
}

}  // namespace tensorflow