1// Copyright (c) 2011 Google, Inc. 2// 3// Permission is hereby granted, free of charge, to any person obtaining a copy 4// of this software and associated documentation files (the "Software"), to deal 5// in the Software without restriction, including without limitation the rights 6// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell 7// copies of the Software, and to permit persons to whom the Software is 8// furnished to do so, subject to the following conditions: 9// 10// The above copyright notice and this permission notice shall be included in 11// all copies or substantial portions of the Software. 12// 13// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR 14// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, 15// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE 16// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER 17// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, 18// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN 19// THE SOFTWARE. 20// 21// CityHash, by Geoff Pike and Jyrki Alakuijala 22// 23// This file provides a few functions for hashing strings. On x86-64 24// hardware in 2011, CityHash64() is faster than other high-quality 25// hash functions, such as Murmur. This is largely due to higher 26// instruction-level parallelism. CityHash64() and CityHash128() also perform 27// well on hash-quality tests. 28// 29// CityHash128() is optimized for relatively long strings and returns 30// a 128-bit hash. For strings more than about 2000 bytes it can be 31// faster than CityHash64(). 32// 33// Functions in the CityHash family are not suitable for cryptography. 34// 35// WARNING: This code has not been tested on big-endian platforms! 36// It is known to work well on little-endian platforms that have a small penalty 37// for unaligned reads, such as current Intel and AMD moderate-to-high-end CPUs. 38// 39// By the way, for some hash functions, given strings a and b, the hash 40// of a+b is easily derived from the hashes of a and b. This property 41// doesn't hold for any hash functions in this file. 42 43#ifndef CITY_HASH_H_ 44#define CITY_HASH_H_ 45 46#include "Platform.h" 47#include <stdlib.h> // for size_t. 48//#include <stdint.h> 49#include <utility> 50 51typedef uint8_t uint8; 52typedef uint32_t uint32; 53typedef uint64_t uint64; 54typedef std::pair<uint64, uint64> uint128; 55 56inline uint64 Uint128Low64(const uint128& x) { return x.first; } 57inline uint64 Uint128High64(const uint128& x) { return x.second; } 58 59// Hash function for a byte array. 60uint64 CityHash64(const char *buf, size_t len); 61 62// Hash function for a byte array. For convenience, a 64-bit seed is also 63// hashed into the result. 64uint64 CityHash64WithSeed(const char *buf, size_t len, uint64 seed); 65 66// Hash function for a byte array. For convenience, two seeds are also 67// hashed into the result. 68uint64 CityHash64WithSeeds(const char *buf, size_t len, 69 uint64 seed0, uint64 seed1); 70 71// Hash function for a byte array. 72uint128 CityHash128(const char *s, size_t len); 73 74// Hash function for a byte array. For convenience, a 128-bit seed is also 75// hashed into the result. 76uint128 CityHash128WithSeed(const char *s, size_t len, uint128 seed); 77 78// Hash 128 input bits down to 64 bits of output. 79// This is intended to be a reasonably good hash function. 80inline uint64 Hash128to64(const uint128& x) { 81 // Murmur-inspired hashing. 82 const uint64 kMul = 0x9ddfea08eb382d69ULL; 83 uint64 a = (Uint128Low64(x) ^ Uint128High64(x)) * kMul; 84 a ^= (a >> 47); 85 uint64 b = (Uint128High64(x) ^ a) * kMul; 86 b ^= (b >> 47); 87 b *= kMul; 88 return b; 89} 90 91// Conditionally include declarations for versions of City that require SSE4.2 92// instructions to be available. 93#if defined(__SSE4_2__) && defined(__x86_64__) 94 95// Hash function for a byte array. 96uint128 CityHashCrc128(const char *s, size_t len); 97 98// Hash function for a byte array. For convenience, a 128-bit seed is also 99// hashed into the result. 100uint128 CityHashCrc128WithSeed(const char *s, size_t len, uint128 seed); 101 102// Hash function for a byte array. Sets result[0] ... result[3]. 103void CityHashCrc256(const char *s, size_t len, uint64 *result); 104 105#endif // __SSE4_2__ 106 107#endif // CITY_HASH_H_ 108