-
Notifications
You must be signed in to change notification settings - Fork 8.7k
/
primitives.h
295 lines (271 loc) · 7.25 KB
/
primitives.h
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
/*
* Licensed to the Apache Software Foundation (ASF) under one
* or more contributor license agreements. See the NOTICE file
* distributed with this work for additional information
* regarding copyright ownership. The ASF licenses this file
* to you under the Apache License, Version 2.0 (the
* "License"); you may not use this file except in compliance
* with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
/**
* High performance primitive functions
*
**/
#ifndef PRIMITIVES_H_
#define PRIMITIVES_H_
#include <stddef.h>
#include <stdint.h>
#include <assert.h>
#include <string>
#ifdef __GNUC__
#define likely(x) __builtin_expect((x),1)
#define unlikely(x) __builtin_expect((x),0)
#else
#define likely(x) (x)
#define unlikely(x) (x)
#endif
//#define SIMPLE_MEMCPY
#if !defined(SIMPLE_MEMCPY)
#define simple_memcpy memcpy
#define simple_memcpy2 memcpy
#else
/**
* This memcpy assumes src & dest are not overlapped,
* and len are normally very small(<64)
* This function is primarily optimized for x86-64 processors,
* on which unaligned 64-bit loads and stores are cheap
*
* @param dest: dest buffer
* @param src: src buffer
* @param len: src buffer size, must be >0
*/
inline void simple_memcpy(void * dest, const void * src, size_t len) {
const uint8_t * src8 = (const uint8_t*)src;
uint8_t * dest8 = (uint8_t*)dest;
switch (len) {
case 0:
return;
case 1:
dest8[0]=src8[0];
return;
case 2:
*(uint16_t*)dest8=*(const uint16_t*)src8;
return;
case 3:
*(uint16_t*)dest8 = *(const uint16_t*)src8;
dest8[2]=src8[2];
return;
case 4:
*(uint32_t*)dest8 = *(const uint32_t*)src8;
return;
}
if (len<8) {
*(uint32_t*)dest8 = *(const uint32_t*)src8;
*(uint32_t*)(dest8+len-4) = *(const uint32_t*)(src8+len-4);
return;
}
if (len<128) {
int64_t cur = (int64_t)len - 8;
while (cur>0) {
*(uint64_t*)(dest8+cur) = *(const uint64_t*)(src8+cur);
cur -= 8;
}
*(uint64_t*)(dest8) = *(const uint64_t*)(src8);
return;
}
::memcpy(dest, src, len);
}
#endif
/**
* little-endian to big-endian or vice versa
*/
inline uint32_t bswap(uint32_t val) {
#ifdef __aarch64__
__asm__("rev %w[dst], %w[src]" : [dst]"=r"(val) : [src]"r"(val));
#else
__asm__("bswap %0" : "=r" (val) : "0" (val));
#endif
return val;
}
inline uint64_t bswap64(uint64_t val) {
#ifdef __aarch64__
__asm__("rev %[dst], %[src]" : [dst]"=r"(val) : [src]"r"(val));
#else
#ifdef __X64
__asm__("bswapq %0" : "=r" (val) : "0" (val));
#else
uint64_t lower = val & 0xffffffffU;
uint32_t higher = (val >> 32) & 0xffffffffU;
lower = bswap(lower);
higher = bswap(higher);
return (lower << 32) + higher;
#endif
#endif
return val;
}
/**
* Fast memcmp
*/
inline int64_t fmemcmp(const char * src, const char * dest, uint32_t len) {
#ifdef BUILDIN_MEMCMP
return memcmp(src, dest, len);
#else
const uint8_t * src8 = (const uint8_t*)src;
const uint8_t * dest8 = (const uint8_t*)dest;
switch (len) {
case 0:
return 0;
case 1:
return (int64_t)src8[0] - (int64_t)dest8[0];
case 2: {
int64_t ret = ((int64_t)src8[0] - (int64_t)dest8[0]);
if (ret)
return ret;
return ((int64_t)src8[1] - (int64_t)dest8[1]);
}
case 3: {
int64_t ret = ((int64_t)src8[0] - (int64_t)dest8[0]);
if (ret)
return ret;
ret = ((int64_t)src8[1] - (int64_t)dest8[1]);
if (ret)
return ret;
return ((int64_t)src8[2] - (int64_t)dest8[2]);
}
case 4: {
return (int64_t)bswap(*(uint32_t*)src) - (int64_t)bswap(*(uint32_t*)dest);
}
}
if (len < 8) {
int64_t ret = ((int64_t)bswap(*(uint32_t*)src) - (int64_t)bswap(*(uint32_t*)dest));
if (ret) {
return ret;
}
return ((int64_t)bswap(*(uint32_t*)(src + len - 4))
- (int64_t)bswap(*(uint32_t*)(dest + len - 4)));
}
uint32_t cur = 0;
uint32_t end = len & (0xffffffffU << 3);
while (cur < end) {
uint64_t l = *(uint64_t*)(src8 + cur);
uint64_t r = *(uint64_t*)(dest8 + cur);
if (l != r) {
l = bswap64(l);
r = bswap64(r);
return l > r ? 1 : -1;
}
cur += 8;
}
uint64_t l = *(uint64_t*)(src8 + len - 8);
uint64_t r = *(uint64_t*)(dest8 + len - 8);
if (l != r) {
l = bswap64(l);
r = bswap64(r);
return l > r ? 1 : -1;
}
return 0;
#endif
}
inline int64_t fmemcmp(const char * src, const char * dest, uint32_t srcLen, uint32_t destLen) {
uint32_t minlen = srcLen < destLen ? srcLen : destLen;
int64_t ret = fmemcmp(src, dest, minlen);
if (ret) {
return ret;
}
return (int64_t)srcLen - (int64_t)destLen;
}
/**
* Fast memory equal
*/
inline bool fmemeq(const char * src, const char * dest, uint32_t len) {
#ifdef BUILDIN_MEMCMP
return 0 == memcmp(src, dest, len);
#else
const uint8_t * src8 = (const uint8_t*)src;
const uint8_t * dest8 = (const uint8_t*)dest;
switch (len) {
case 0:
return true;
case 1:
return src8[0] == dest8[0];
case 2:
return *(uint16_t*)src8 == *(uint16_t*)dest8;
case 3:
return (*(uint16_t*)src8 == *(uint16_t*)dest8) && (src8[2] == dest8[2]);
case 4:
return *(uint32_t*)src8 == *(uint32_t*)dest8;
}
if (len < 8) {
return (*(uint32_t*)src8 == *(uint32_t*)dest8)
&& (*(uint32_t*)(src8 + len - 4) == *(uint32_t*)(dest8 + len - 4));
}
uint32_t cur = 0;
uint32_t end = len & (0xffffffff << 3);
while (cur < end) {
uint64_t l = *(uint64_t*)(src8 + cur);
uint64_t r = *(uint64_t*)(dest8 + cur);
if (l != r) {
return false;
}
cur += 8;
}
uint64_t l = *(uint64_t*)(src8 + len - 8);
uint64_t r = *(uint64_t*)(dest8 + len - 8);
if (l != r) {
return false;
}
return true;
#endif
}
inline bool fmemeq(const char * src, uint32_t srcLen, const char * dest, uint32_t destLen) {
if (srcLen != destLen) {
return false;
}
return fmemeq(src, dest, std::min(srcLen, destLen));
}
/**
* Fast memory equal, reverse order
*/
inline bool frmemeq(const char * src, const char * dest, uint32_t len) {
const uint8_t * src8 = (const uint8_t*)src;
const uint8_t * dest8 = (const uint8_t*)dest;
switch (len) {
case 0:
return true;
case 1:
return src8[0] == dest8[0];
case 2:
return *(uint16_t*)src8 == *(uint16_t*)dest8;
case 3:
return (src8[2] == dest8[2]) && (*(uint16_t*)src8 == *(uint16_t*)dest8);
case 4:
return *(uint32_t*)src8 == *(uint32_t*)dest8;
}
if (len < 8) {
return (*(uint32_t*)(src8 + len - 4) == *(uint32_t*)(dest8 + len - 4))
&& (*(uint32_t*)src8 == *(uint32_t*)dest8);
}
int32_t cur = (int32_t)len - 8;
while (cur > 0) {
if (*(uint64_t*)(src8 + cur) != *(uint64_t*)(dest8 + cur)) {
return false;
}
cur -= 8;
}
return *(uint64_t*)(src8) == *(uint64_t*)(dest8);
}
inline bool frmemeq(const char * src, const char * dest, uint32_t srcLen, uint32_t destLen) {
if (srcLen != destLen) {
return false;
}
return frmemeq(src, dest, std::min(srcLen, destLen));
}
#endif /* PRIMITIVES_H_ */