murmur.cpp 2.2 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111
  1. /*
  2. * Copyright (c) 2012-2014 Daniele Bartolini and individual contributors.
  3. * License: https://github.com/taylor001/crown/blob/master/LICENSE
  4. */
  5. #include "murmur.h"
  6. namespace crown
  7. {
  8. /// MurmurHash2, by Austin Appleby
  9. ///
  10. /// @note
  11. /// This code makes a few assumptions about how your machine behaves
  12. ///
  13. /// 1. We can read a 4-byte value from any address without crashing
  14. /// 2. sizeof(int) == 4
  15. ///
  16. /// And it has a few limitations -
  17. ///
  18. /// 1. It will not work incrementally.
  19. /// 2. It will not produce the same results on little-endian and big-endian
  20. /// machines.
  21. uint32_t murmur2_32(const void* key, size_t len, uint32_t seed)
  22. {
  23. // 'm' and 'r' are mixing constants generated offline.
  24. // They're not really 'magic', they just happen to work well.
  25. const unsigned int m = 0x5bd1e995;
  26. const int r = 24;
  27. // Initialize the hash to a 'random' value
  28. unsigned int h = seed ^ len;
  29. // Mix 4 bytes at a time into the hash
  30. const unsigned char * data = (const unsigned char *)key;
  31. while(len >= 4)
  32. {
  33. unsigned int k = *(unsigned int *)data;
  34. k *= m;
  35. k ^= k >> r;
  36. k *= m;
  37. h *= m;
  38. h ^= k;
  39. data += 4;
  40. len -= 4;
  41. }
  42. // Handle the last few bytes of the input array
  43. switch(len)
  44. {
  45. case 3: h ^= data[2] << 16;
  46. case 2: h ^= data[1] << 8;
  47. case 1: h ^= data[0];
  48. h *= m;
  49. };
  50. // Do a few final mixes of the hash to ensure the last few
  51. // bytes are well-incorporated.
  52. h ^= h >> 13;
  53. h *= m;
  54. h ^= h >> 15;
  55. return h;
  56. }
  57. uint64_t murmur2_64(const void* key, int len, uint64_t seed)
  58. {
  59. const uint64_t m = 0xc6a4a7935bd1e995ull;
  60. const int r = 47;
  61. uint64_t h = seed ^ (len * m);
  62. const uint64_t * data = (const uint64_t *)key;
  63. const uint64_t * end = data + (len/8);
  64. while(data != end)
  65. {
  66. uint64_t k = *data++;
  67. k *= m;
  68. k ^= k >> r;
  69. k *= m;
  70. h ^= k;
  71. h *= m;
  72. }
  73. const unsigned char * data2 = (const unsigned char*)data;
  74. switch(len & 7)
  75. {
  76. case 7: h ^= uint64_t(data2[6]) << 48;
  77. case 6: h ^= uint64_t(data2[5]) << 40;
  78. case 5: h ^= uint64_t(data2[4]) << 32;
  79. case 4: h ^= uint64_t(data2[3]) << 24;
  80. case 3: h ^= uint64_t(data2[2]) << 16;
  81. case 2: h ^= uint64_t(data2[1]) << 8;
  82. case 1: h ^= uint64_t(data2[0]);
  83. h *= m;
  84. };
  85. h ^= h >> r;
  86. h *= m;
  87. h ^= h >> r;
  88. return h;
  89. }
  90. } // namespace crown