OLD | NEW |
1 /* | 1 /* |
2 * Copyright 2012 Google Inc. | 2 * Copyright 2012 Google Inc. |
3 * | 3 * |
4 * Use of this source code is governed by a BSD-style license that can be | 4 * Use of this source code is governed by a BSD-style license that can be |
5 * found in the LICENSE file. | 5 * found in the LICENSE file. |
6 */ | 6 */ |
7 | 7 |
8 #ifndef SkChecksum_DEFINED | 8 #ifndef SkChecksum_DEFINED |
9 #define SkChecksum_DEFINED | 9 #define SkChecksum_DEFINED |
10 | 10 |
| 11 #include "SkString.h" |
| 12 #include "SkTLogic.h" |
11 #include "SkTypes.h" | 13 #include "SkTypes.h" |
12 | 14 |
13 /** | 15 /** |
14 * Computes a 32bit checksum from a blob of 32bit aligned data. This is meant | 16 * Computes a 32bit checksum from a blob of 32bit aligned data. This is meant |
15 * to be very very fast, as it is used internally by the font cache, in | 17 * to be very very fast, as it is used internally by the font cache, in |
16 * conjuction with the entire raw key. This algorithm does not generate | 18 * conjuction with the entire raw key. This algorithm does not generate |
17 * unique values as well as others (e.g. MD5) but it performs much faster. | 19 * unique values as well as others (e.g. MD5) but it performs much faster. |
18 * Skia's use cases can survive non-unique values (since the entire key is | 20 * Skia's use cases can survive non-unique values (since the entire key is |
19 * always available). Clients should only be used in circumstances where speed | 21 * always available). Clients should only be used in circumstances where speed |
20 * over uniqueness is at a premium. | 22 * over uniqueness is at a premium. |
(...skipping 41 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
62 hash *= 0x85ebca6b; | 64 hash *= 0x85ebca6b; |
63 hash ^= hash >> 16; | 65 hash ^= hash >> 16; |
64 return hash; | 66 return hash; |
65 } | 67 } |
66 | 68 |
67 /** | 69 /** |
68 * Calculate 32-bit Murmur hash (murmur3). | 70 * Calculate 32-bit Murmur hash (murmur3). |
69 * This should take 2-3x longer than SkChecksum::Compute, but is a considera
bly better hash. | 71 * This should take 2-3x longer than SkChecksum::Compute, but is a considera
bly better hash. |
70 * See en.wikipedia.org/wiki/MurmurHash. | 72 * See en.wikipedia.org/wiki/MurmurHash. |
71 * | 73 * |
72 * @param data Memory address of the data block to be processed. Must be 32
-bit aligned. | 74 * @param data Memory address of the data block to be processed. |
73 * @param size Size of the data block in bytes. Must be a multiple of 4. | 75 * @param size Size of the data block in bytes. |
74 * @param seed Initial hash seed. (optional) | 76 * @param seed Initial hash seed. (optional) |
75 * @return hash result | 77 * @return hash result |
76 */ | 78 */ |
77 static uint32_t Murmur3(const uint32_t* data, size_t bytes, uint32_t seed=0)
{ | 79 static uint32_t Murmur3(const void* data, size_t bytes, uint32_t seed=0) { |
78 // Use may_alias to remind the compiler we're intentionally violating st
rict aliasing, | 80 // Use may_alias to remind the compiler we're intentionally violating st
rict aliasing, |
79 // and so not to apply strict-aliasing-based optimizations. | 81 // and so not to apply strict-aliasing-based optimizations. |
80 typedef uint32_t SK_ATTRIBUTE(may_alias) aliased_uint32_t; | 82 typedef uint32_t SK_ATTRIBUTE(may_alias) aliased_uint32_t; |
| 83 typedef uint8_t SK_ATTRIBUTE(may_alias) aliased_uint8_t; |
| 84 |
| 85 // Handle 4 bytes at a time while possible. |
81 const aliased_uint32_t* safe_data = (const aliased_uint32_t*)data; | 86 const aliased_uint32_t* safe_data = (const aliased_uint32_t*)data; |
82 | |
83 SkASSERTF(SkIsAlign4(bytes), "Expected 4-byte multiple, got %zu", bytes)
; | |
84 const size_t words = bytes/4; | 87 const size_t words = bytes/4; |
85 | |
86 | |
87 uint32_t hash = seed; | 88 uint32_t hash = seed; |
88 for (size_t i = 0; i < words; i++) { | 89 for (size_t i = 0; i < words; i++) { |
89 uint32_t k = safe_data[i]; | 90 uint32_t k = safe_data[i]; |
90 k *= 0xcc9e2d51; | 91 k *= 0xcc9e2d51; |
91 k = (k << 15) | (k >> 17); | 92 k = (k << 15) | (k >> 17); |
92 k *= 0x1b873593; | 93 k *= 0x1b873593; |
93 | 94 |
94 hash ^= k; | 95 hash ^= k; |
95 hash = (hash << 13) | (hash >> 19); | 96 hash = (hash << 13) | (hash >> 19); |
96 hash *= 5; | 97 hash *= 5; |
97 hash += 0xe6546b64; | 98 hash += 0xe6546b64; |
98 } | 99 } |
| 100 |
| 101 // Handle last 0-3 bytes. |
| 102 const aliased_uint8_t* safe_tail = (const uint8_t*)(safe_data + words); |
| 103 uint32_t k = 0; |
| 104 switch (bytes & 3) { |
| 105 case 3: k ^= safe_tail[2] << 16; |
| 106 case 2: k ^= safe_tail[1] << 8; |
| 107 case 1: k ^= safe_tail[0] << 0; |
| 108 k *= 0xcc9e2d51; |
| 109 k = (k << 15) | (k >> 17); |
| 110 k *= 0x1b873593; |
| 111 hash ^= k; |
| 112 } |
| 113 |
99 hash ^= bytes; | 114 hash ^= bytes; |
100 return Mix(hash); | 115 return Mix(hash); |
101 } | 116 } |
102 | 117 |
103 /** | 118 /** |
104 * Compute a 32-bit checksum for a given data block | 119 * Compute a 32-bit checksum for a given data block |
105 * | 120 * |
106 * WARNING: this algorithm is tuned for efficiency, not backward/forward | 121 * WARNING: this algorithm is tuned for efficiency, not backward/forward |
107 * compatibility. It may change at any time, so a checksum generated with | 122 * compatibility. It may change at any time, so a checksum generated with |
108 * one version of the Skia code may not match a checksum generated with | 123 * one version of the Skia code may not match a checksum generated with |
(...skipping 49 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
158 * since >>32 is undefined for 32bit ints, hence the wacky HALFBITS | 173 * since >>32 is undefined for 32bit ints, hence the wacky HALFBITS |
159 * define. | 174 * define. |
160 */ | 175 */ |
161 if (8 == sizeof(result)) { | 176 if (8 == sizeof(result)) { |
162 result ^= result >> HALFBITS; | 177 result ^= result >> HALFBITS; |
163 } | 178 } |
164 return static_cast<uint32_t>(result); | 179 return static_cast<uint32_t>(result); |
165 } | 180 } |
166 }; | 181 }; |
167 | 182 |
| 183 // SkGoodHash should usually be your first choice in hashing data. |
| 184 // It should be both reasonably fast and high quality. |
| 185 |
| 186 template <typename K> |
| 187 uint32_t SkGoodHash(const K& k) { |
| 188 if (sizeof(K) == 4) { |
| 189 return SkChecksum::Mix(*(const uint32_t*)&k); |
| 190 } |
| 191 return SkChecksum::Murmur3(&k, sizeof(K)); |
| 192 } |
| 193 |
| 194 inline uint32_t SkGoodHash(const SkString& k) { |
| 195 return SkChecksum::Murmur3(k.c_str(), k.size()); |
| 196 } |
| 197 |
168 #endif | 198 #endif |
OLD | NEW |