summaryrefslogtreecommitdiff
path: root/MurmurHash1.cpp
blob: 82255667c376b064e7c813b7a5422d515a3f17ba (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
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
//-----------------------------------------------------------------------------
// MurmurHash was written by Austin Appleby, and is placed in the public
// domain. The author hereby disclaims copyright to this source code.

// Note - This code makes a few assumptions about how your machine behaves -

// 1. We can read a 4-byte value from any address without crashing
// 2. sizeof(int) == 4

// And it has a few limitations -

// 1. It will not work incrementally.
// 2. It will not produce the same results on little-endian and big-endian
//    machines.

#include "MurmurHash1.h"

//-----------------------------------------------------------------------------

uint32_t MurmurHash1 ( const void * key, int len, uint32_t seed )
{
  const unsigned int m = 0xc6a4a793;

  const int r = 16;

  unsigned int h = seed ^ (len * m);

  //----------
  
  const unsigned char * data = (const unsigned char *)key;

  while(len >= 4)
  {
    unsigned int k = *(unsigned int *)data;

    h += k;
    h *= m;
    h ^= h >> 16;

    data += 4;
    len -= 4;
  }
  
  //----------
  
  switch(len)
  {
  case 3:
    h += data[2] << 16;
  case 2:
    h += data[1] << 8;
  case 1:
    h += data[0];
    h *= m;
    h ^= h >> r;
  };
 
  //----------

  h *= m;
  h ^= h >> 10;
  h *= m;
  h ^= h >> 17;

  return h;
} 

//-----------------------------------------------------------------------------
// MurmurHash1Aligned, by Austin Appleby

// Same algorithm as MurmurHash1, but only does aligned reads - should be safer
// on certain platforms. 

// Performance should be equal to or better than the simple version.

unsigned int MurmurHash1Aligned ( const void * key, int len, unsigned int seed )
{
  const unsigned int m = 0xc6a4a793;
  const int r = 16;

  const unsigned char * data = (const unsigned char *)key;

  unsigned int h = seed ^ (len * m);

  int align = (uint64_t)data & 3;

  if(align && (len >= 4))
  {
    // Pre-load the temp registers

    unsigned int t = 0, d = 0;

    switch(align)
    {
      case 1: t |= data[2] << 16;
      case 2: t |= data[1] << 8;
      case 3: t |= data[0];
    }

    t <<= (8 * align);

    data += 4-align;
    len -= 4-align;

    int sl = 8 * (4-align);
    int sr = 8 * align;

    // Mix

    while(len >= 4)
    {
      d = *(unsigned int *)data;
      t = (t >> sr) | (d << sl);
      h += t;
      h *= m;
      h ^= h >> r;
      t = d;

      data += 4;
      len -= 4;
    }

    // Handle leftover data in temp registers

    int pack = len < align ? len : align;

    d = 0;

    switch(pack)
    {
    case 3: d |= data[2] << 16;
    case 2: d |= data[1] << 8;
    case 1: d |= data[0];
    case 0: h += (t >> sr) | (d << sl);
        h *= m;
        h ^= h >> r;
    }

    data += pack;
    len -= pack;
  }
  else
  {
    while(len >= 4)
    {
      h += *(unsigned int *)data;
      h *= m;
      h ^= h >> r;

      data += 4;
      len -= 4;
    }
  }

  //----------
  // Handle tail bytes

  switch(len)
  {
  case 3: h += data[2] << 16;
  case 2: h += data[1] << 8;
  case 1: h += data[0];
      h *= m;
      h ^= h >> r;
  };

  h *= m;
  h ^= h >> 10;
  h *= m;
  h ^= h >> 17;

  return h;
}