File size: 13,818 Bytes
9375c9a |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 |
// Copyright (C) 2011 Davis E. King ([email protected])
// License: Boost Software License See LICENSE.txt for the full license.
#ifndef DLIB_MURMUR_HAsH_3_Hh_
#define DLIB_MURMUR_HAsH_3_Hh_
#include "murmur_hash3_abstract.h"
#include "../uintn.h"
#include <utility>
#include <string.h>
namespace dlib
{
//-----------------------------------------------------------------------------
// The original MurmurHash3 code was written by Austin Appleby, and is placed
// in the public domain. The author hereby disclaims copyright to this source code.
// The code in this particular file was modified by Davis E. King. In
// particular, endian-swapping was added along with some other minor code
// changes like avoiding strict aliasing violations.
//-----------------------------------------------------------------------------
// Platform-specific functions and macros
// Microsoft Visual Studio
#if ((defined(__GNUC__) && __GNUC__ >= 7) || defined(__clang__))
#define DLIB_FALLTHROUGH [[fallthrough]]
#else
#define DLIB_FALLTHROUGH
#endif
#if defined(_MSC_VER)
#define DLIB_FORCE_INLINE __forceinline
#include <stdlib.h>
#define DLIB_ROTL32(x,y) _rotl(x,y)
#define DLIB_ROTL64(x,y) _rotl64(x,y)
#define DLIB_BIG_CONSTANT(x) (x)
// Other compilers
#else // defined(_MSC_VER)
#define DLIB_FORCE_INLINE __attribute__((always_inline)) inline
inline uint32 murmur_rotl32 ( uint32 x, int8 r )
{
return (x << r) | (x >> (32 - r));
}
inline uint64 murmur_rotl64 ( uint64 x, int8 r )
{
return (x << r) | (x >> (64 - r));
}
#define DLIB_ROTL32(x,y) dlib::murmur_rotl32(x,y)
#define DLIB_ROTL64(x,y) dlib::murmur_rotl64(x,y)
#define DLIB_BIG_CONSTANT(x) (x##LLU)
#endif // !defined(_MSC_VER)
// ----------------------------------------------------------------------------------------
// Block read - if your platform needs to do endian-swapping or can only
// handle aligned reads, do the conversion here
DLIB_FORCE_INLINE uint32 murmur_getblock ( const uint32 * p, int i )
{
// The reason we do a memcpy() here instead of simply returning p[i] is because
// doing it this way avoids violations of the strict aliasing rule when all these
// functions are inlined into the user's code.
uint32 temp;
memcpy(&temp, p+i, 4);
return temp;
}
DLIB_FORCE_INLINE uint32 murmur_getblock_byte_swap ( const uint32 * p, int i )
{
union
{
uint8 bytes[4];
uint32 val;
} temp;
const uint8* pp = reinterpret_cast<const uint8*>(p + i);
temp.bytes[0] = pp[3];
temp.bytes[1] = pp[2];
temp.bytes[2] = pp[1];
temp.bytes[3] = pp[0];
return temp.val;
}
DLIB_FORCE_INLINE uint64 murmur_getblock ( const uint64 * p, int i )
{
// The reason we do a memcpy() here instead of simply returning p[i] is because
// doing it this way avoids violations of the strict aliasing rule when all these
// functions are inlined into the user's code.
uint64 temp;
memcpy(&temp, p+i, 8);
return temp;
}
DLIB_FORCE_INLINE uint64 murmur_getblock_byte_swap ( const uint64 * p, int i )
{
union
{
uint8 bytes[8];
uint64 val;
} temp;
const uint8* pp = reinterpret_cast<const uint8*>(p + i);
temp.bytes[0] = pp[7];
temp.bytes[1] = pp[6];
temp.bytes[2] = pp[5];
temp.bytes[3] = pp[4];
temp.bytes[4] = pp[3];
temp.bytes[5] = pp[2];
temp.bytes[6] = pp[1];
temp.bytes[7] = pp[0];
return temp.val;
}
// ----------------------------------------------------------------------------------------
// Finalization mix - force all bits of a hash block to avalanche
DLIB_FORCE_INLINE uint32 murmur_fmix ( uint32 h )
{
h ^= h >> 16;
h *= 0x85ebca6b;
h ^= h >> 13;
h *= 0xc2b2ae35;
h ^= h >> 16;
return h;
}
// ----------------------------------------------------------------------------------------
DLIB_FORCE_INLINE uint64 murmur_fmix ( uint64 k )
{
k ^= k >> 33;
k *= DLIB_BIG_CONSTANT(0xff51afd7ed558ccd);
k ^= k >> 33;
k *= DLIB_BIG_CONSTANT(0xc4ceb9fe1a85ec53);
k ^= k >> 33;
return k;
}
// ----------------------------------------------------------------------------------------
inline uint32 murmur_hash3 (
const void * key,
const int len,
const uint32 seed = 0
)
{
const uint8 * data = (const uint8*)key;
const int nblocks = len / 4;
uint32 h1 = seed;
uint32 c1 = 0xcc9e2d51;
uint32 c2 = 0x1b873593;
//----------
// body
const uint32 * blocks = (const uint32 *)(data + nblocks*4);
bool is_little_endian = true;
uint32 endian_test = 1;
if (*reinterpret_cast<unsigned char*>(&endian_test) != 1)
is_little_endian = false;
if (is_little_endian)
{
for(int i = -nblocks; i; i++)
{
uint32 k1 = murmur_getblock(blocks,i);
k1 *= c1;
k1 = DLIB_ROTL32(k1,15);
k1 *= c2;
h1 ^= k1;
h1 = DLIB_ROTL32(h1,13);
h1 = h1*5+0xe6546b64;
}
}
else
{
for(int i = -nblocks; i; i++)
{
uint32 k1 = murmur_getblock_byte_swap(blocks,i);
k1 *= c1;
k1 = DLIB_ROTL32(k1,15);
k1 *= c2;
h1 ^= k1;
h1 = DLIB_ROTL32(h1,13);
h1 = h1*5+0xe6546b64;
}
}
//----------
// tail
const uint8 * tail = (const uint8*)(data + nblocks*4);
uint32 k1 = 0;
switch(len & 3)
{
case 3: k1 ^= tail[2] << 16;
DLIB_FALLTHROUGH; // fall through
case 2: k1 ^= tail[1] << 8;
DLIB_FALLTHROUGH; // fall through
case 1: k1 ^= tail[0];
k1 *= c1; k1 = DLIB_ROTL32(k1,15); k1 *= c2; h1 ^= k1;
};
//----------
// finalization
h1 ^= len;
h1 = murmur_fmix(h1);
return h1;
}
// ----------------------------------------------------------------------------------------
inline uint32 murmur_hash3_2 (
const uint32 v1,
const uint32 v2
)
{
uint32 h1 = v2;
uint32 c1 = 0xcc9e2d51;
uint32 c2 = 0x1b873593;
//----------
// body
uint32 k1 = v1;
k1 *= c1;
k1 = DLIB_ROTL32(k1,15);
k1 *= c2;
h1 ^= k1;
h1 = DLIB_ROTL32(h1,13);
h1 = h1*5+0xe6546b64;
//----------
// finalization
h1 ^= 4; // =^ by length in bytes
h1 = murmur_fmix(h1);
return h1;
}
// ----------------------------------------------------------------------------------------
inline uint32 murmur_hash3_3 (
const uint32 v1,
const uint32 v2,
const uint32 v3
)
{
uint32 h1 = v3;
uint32 c1 = 0xcc9e2d51;
uint32 c2 = 0x1b873593;
//----------
// body
uint32 k1 = v1;
k1 *= c1;
k1 = DLIB_ROTL32(k1,15);
k1 *= c2;
h1 ^= k1;
h1 = DLIB_ROTL32(h1,13);
h1 = h1*5+0xe6546b64;
k1 = v2;
k1 *= c1;
k1 = DLIB_ROTL32(k1,15);
k1 *= c2;
h1 ^= k1;
h1 = DLIB_ROTL32(h1,13);
h1 = h1*5+0xe6546b64;
//----------
// finalization
h1 ^= 8; // =^ by length in bytes
h1 = murmur_fmix(h1);
return h1;
}
// ----------------------------------------------------------------------------------------
inline std::pair<uint64,uint64> murmur_hash3_128bit (
const void* key,
const int len,
const uint64 seed = 0
)
{
const uint8 * data = (const uint8*)key;
const int nblocks = len / 16;
uint64 h1 = seed;
uint64 h2 = seed;
uint64 c1 = DLIB_BIG_CONSTANT(0x87c37b91114253d5);
uint64 c2 = DLIB_BIG_CONSTANT(0x4cf5ad432745937f);
//----------
// body
const uint64 * blocks = (const uint64 *)(data);
bool is_little_endian = true;
uint32 endian_test = 1;
if (*reinterpret_cast<unsigned char*>(&endian_test) != 1)
is_little_endian = false;
if (is_little_endian)
{
for(int i = 0; i < nblocks; i++)
{
uint64 k1 = murmur_getblock(blocks,i*2+0);
uint64 k2 = murmur_getblock(blocks,i*2+1);
k1 *= c1; k1 = DLIB_ROTL64(k1,31); k1 *= c2; h1 ^= k1;
h1 = DLIB_ROTL64(h1,27); h1 += h2; h1 = h1*5+0x52dce729;
k2 *= c2; k2 = DLIB_ROTL64(k2,33); k2 *= c1; h2 ^= k2;
h2 = DLIB_ROTL64(h2,31); h2 += h1; h2 = h2*5+0x38495ab5;
}
}
else
{
for(int i = 0; i < nblocks; i++)
{
uint64 k1 = murmur_getblock_byte_swap(blocks,i*2+0);
uint64 k2 = murmur_getblock_byte_swap(blocks,i*2+1);
k1 *= c1; k1 = DLIB_ROTL64(k1,31); k1 *= c2; h1 ^= k1;
h1 = DLIB_ROTL64(h1,27); h1 += h2; h1 = h1*5+0x52dce729;
k2 *= c2; k2 = DLIB_ROTL64(k2,33); k2 *= c1; h2 ^= k2;
h2 = DLIB_ROTL64(h2,31); h2 += h1; h2 = h2*5+0x38495ab5;
}
}
//----------
// tail
const uint8 * tail = (const uint8*)(data + nblocks*16);
uint64 k1 = 0;
uint64 k2 = 0;
switch(len & 15)
{
case 15: k2 ^= uint64(tail[14]) << 48; DLIB_FALLTHROUGH; // fall through
case 14: k2 ^= uint64(tail[13]) << 40; DLIB_FALLTHROUGH; // fall through
case 13: k2 ^= uint64(tail[12]) << 32; DLIB_FALLTHROUGH; // fall through
case 12: k2 ^= uint64(tail[11]) << 24; DLIB_FALLTHROUGH; // fall through
case 11: k2 ^= uint64(tail[10]) << 16; DLIB_FALLTHROUGH; // fall through
case 10: k2 ^= uint64(tail[ 9]) << 8; DLIB_FALLTHROUGH; // fall through
case 9: k2 ^= uint64(tail[ 8]) << 0;
k2 *= c2; k2 = DLIB_ROTL64(k2,33); k2 *= c1; h2 ^= k2; DLIB_FALLTHROUGH; // fall through
case 8: k1 ^= uint64(tail[ 7]) << 56; DLIB_FALLTHROUGH; // fall through
case 7: k1 ^= uint64(tail[ 6]) << 48; DLIB_FALLTHROUGH; // fall through
case 6: k1 ^= uint64(tail[ 5]) << 40; DLIB_FALLTHROUGH; // fall through
case 5: k1 ^= uint64(tail[ 4]) << 32; DLIB_FALLTHROUGH; // fall through
case 4: k1 ^= uint64(tail[ 3]) << 24; DLIB_FALLTHROUGH; // fall through
case 3: k1 ^= uint64(tail[ 2]) << 16; DLIB_FALLTHROUGH; // fall through
case 2: k1 ^= uint64(tail[ 1]) << 8; DLIB_FALLTHROUGH; // fall through
case 1: k1 ^= uint64(tail[ 0]) << 0;
k1 *= c1; k1 = DLIB_ROTL64(k1,31); k1 *= c2; h1 ^= k1;
};
//----------
// finalization
h1 ^= len; h2 ^= len;
h1 += h2;
h2 += h1;
h1 = murmur_fmix(h1);
h2 = murmur_fmix(h2);
h1 += h2;
h2 += h1;
return std::make_pair(h1,h2);
}
// ----------------------------------------------------------------------------------------
inline std::pair<uint64,uint64> murmur_hash3_128bit (
const uint32& v1,
const uint32& v2,
const uint32& v3,
const uint32& v4
)
{
uint64 h1 = 0;
uint64 h2 = 0;
const uint64 c1 = DLIB_BIG_CONSTANT(0x87c37b91114253d5);
const uint64 c2 = DLIB_BIG_CONSTANT(0x4cf5ad432745937f);
//----------
// body
uint64 k1 = (static_cast<uint64>(v2)<<32)|v1;
uint64 k2 = (static_cast<uint64>(v4)<<32)|v3;
k1 *= c1; k1 = DLIB_ROTL64(k1,31); k1 *= c2;
h1 = DLIB_ROTL64(k1,27); h1 = h1*5+0x52dce729;
k2 *= c2; k2 = DLIB_ROTL64(k2,33); k2 *= c1;
h2 = DLIB_ROTL64(k2,31); h2 += h1; h2 = h2*5+0x38495ab5;
//----------
// finalization
h1 ^= 16; h2 ^= 16;
h1 += h2;
h2 += h1;
h1 = murmur_fmix(h1);
h2 = murmur_fmix(h2);
h1 += h2;
h2 += h1;
return std::make_pair(h1,h2);
}
// ----------------------------------------------------------------------------------------
inline std::pair<uint64,uint64> murmur_hash3_128bit_3 (
uint64 k1,
uint64 k2,
uint64 k3
)
{
uint64 h1 = k3;
uint64 h2 = k3;
const uint64 c1 = DLIB_BIG_CONSTANT(0x87c37b91114253d5);
const uint64 c2 = DLIB_BIG_CONSTANT(0x4cf5ad432745937f);
//----------
// body
k1 *= c1; k1 = DLIB_ROTL64(k1,31); k1 *= c2; h1 ^= k1;
h1 = DLIB_ROTL64(h1,27); h1 += h2; h1 = h1*5+0x52dce729;
k2 *= c2; k2 = DLIB_ROTL64(k2,33); k2 *= c1; h2 ^= k2;
h2 = DLIB_ROTL64(h2,31); h2 += h1; h2 = h2*5+0x38495ab5;
//----------
// finalization
h1 ^= 16; h2 ^= 16;
h1 += h2;
h2 += h1;
h1 = murmur_fmix(h1);
h2 = murmur_fmix(h2);
h1 += h2;
h2 += h1;
return std::make_pair(h1,h2);
}
// ----------------------------------------------------------------------------------------
}
#endif // DLIB_MURMUR_HAsH_3_Hh_
|