Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Sign up
Appearance settings

Commit 6cc21c4

Browse files
nielsdosdivinity76
andcommitted
Optimize pack()
Instead of using lookup tables, we can use a combination of shifts and byte swapping to achieve the same thing in less cycles and with less code. Benchmark files --------------- pack1.php: ```php for ($i = 0; $i < 10_000_000; ++$i) { pack("J", 0x7FFFFFFFFFFFFFFF); } ``` pack2.php: ```php for ($i = 0; $i < 4000000; ++$i) { pack("nvc*", 0x1234, 0x5678, 65, 66); } ``` On an i7-4790: ``` Benchmark 1: ./sapi/cli/php pack1.php Time (mean ± σ): 408.8 ms ± 3.4 ms [User: 406.1 ms, System: 1.6 ms] Range (min ... max): 403.6 ms ... 413.6 ms 10 runs Benchmark 2: ./sapi/cli/php_old pack1.php Time (mean ± σ): 451.7 ms ± 7.7 ms [User: 448.5 ms, System: 2.0 ms] Range (min ... max): 442.8 ms ... 461.2 ms 10 runs Summary ./sapi/cli/php pack1.php ran 1.11 ± 0.02 times faster than ./sapi/cli/php_old pack1.php Benchmark 1: ./sapi/cli/php pack2.php Time (mean ± σ): 239.3 ms ± 6.0 ms [User: 236.2 ms, System: 2.3 ms] Range (min ... max): 233.2 ms ... 256.8 ms 12 runs Benchmark 2: ./sapi/cli/php_old pack2.php Time (mean ± σ): 271.9 ms ± 3.3 ms [User: 269.7 ms, System: 1.3 ms] Range (min ... max): 267.4 ms ... 279.0 ms 11 runs Summary ./sapi/cli/php pack2.php ran 1.14 ± 0.03 times faster than ./sapi/cli/php_old pack2.php ``` On an i7-1185G7: ``` Benchmark 1: ./sapi/cli/php pack1.php Time (mean ± σ): 263.7 ms ± 1.8 ms [User: 262.6 ms, System: 0.9 ms] Range (min ... max): 261.5 ms ... 268.2 ms 11 runs Benchmark 2: ./sapi/cli/php_old pack1.php Time (mean ± σ): 303.3 ms ± 6.5 ms [User: 300.7 ms, System: 2.3 ms] Range (min ... max): 297.4 ms ... 318.1 ms 10 runs Summary ./sapi/cli/php pack1.php ran 1.15 ± 0.03 times faster than ./sapi/cli/php_old pack1.php Benchmark 1: ./sapi/cli/php pack2.php Time (mean ± σ): 156.7 ms ± 2.9 ms [User: 154.7 ms, System: 1.7 ms] Range (min ... max): 151.6 ms ... 164.7 ms 19 runs Benchmark 2: ./sapi/cli/php_old pack2.php Time (mean ± σ): 174.6 ms ± 3.3 ms [User: 171.9 ms, System: 2.3 ms] Range (min ... max): 170.7 ms ... 180.4 ms 17 runs Summary ./sapi/cli/php pack2.php ran 1.11 ± 0.03 times faster than ./sapi/cli/php_old pack2.php ``` Closes GH-18524. Co-authored-by: divinity76 <divinity76@gmail.com>
1 parent 45215d0 commit 6cc21c4

File tree

6 files changed

+47
-200
lines changed

6 files changed

+47
-200
lines changed

‎NEWS‎

Lines changed: 3 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -2,6 +2,9 @@ PHP NEWS
22
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
33
?? ??? ????, PHP 8.5.0alpha2
44

5+
- Standard:
6+
. Optimized pack(). (nielsdos, divinity76)
7+
58
- URI:
69
. Return the singleton UrlValidationErrorType instances from Uri\WhatWg\Url
710
instead of creating new objects that are different from the singleton.

‎UPGRADING‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -619,6 +619,7 @@ PHP 8.5 UPGRADE NOTES
619619
. Improved performance of urlencode() and rawurlencode().
620620
. Improved unpack() performance with nameless repetitions by avoiding
621621
creating temporary strings and reparsing them.
622+
. Improved pack() performance.
622623

623624
- XMLReader:
624625
. Improved property access performance.

‎ext/standard/basic_functions.c‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -299,7 +299,6 @@ PHP_MINIT_FUNCTION(basic) /* {{{ */
299299

300300
BASIC_MINIT_SUBMODULE(var)
301301
BASIC_MINIT_SUBMODULE(file)
302-
BASIC_MINIT_SUBMODULE(pack)
303302
BASIC_MINIT_SUBMODULE(browscap)
304303
BASIC_MINIT_SUBMODULE(standard_filters)
305304
BASIC_MINIT_SUBMODULE(user_filters)

‎ext/standard/pack.c‎

Lines changed: 43 additions & 176 deletions
Original file line numberDiff line numberDiff line change
@@ -19,7 +19,6 @@
1919
#include <stdlib.h>
2020
#include <errno.h>
2121
#include <sys/types.h>
22-
#include "pack.h"
2322

2423
#define INC_OUTPUTPOS(a,b) \
2524
if ((a) < 0 || ((INT_MAX - outputpos)/((int)b)) < (a)) { \
@@ -30,10 +29,23 @@
3029
} \
3130
outputpos += (a)*(b);
3231

32+
typedef enum {
33+
PHP_LITTLE_ENDIAN,
34+
PHP_BIG_ENDIAN,
35+
} php_pack_endianness;
36+
3337
#ifdef WORDS_BIGENDIAN
34-
#define MACHINE_LITTLE_ENDIAN 0
38+
# define MACHINE_LITTLE_ENDIAN 0
39+
# define PHP_MACHINE_ENDIAN PHP_BIG_ENDIAN
3540
#else
36-
#define MACHINE_LITTLE_ENDIAN 1
41+
# define MACHINE_LITTLE_ENDIAN 1
42+
# define PHP_MACHINE_ENDIAN PHP_LITTLE_ENDIAN
43+
#endif
44+
45+
#ifdef ZEND_ENABLE_ZVAL_LONG64
46+
# define PHP_LONG_BSWAP(u) ZEND_BYTES_SWAP64(u)
47+
#else
48+
# define PHP_LONG_BSWAP(u) ZEND_BYTES_SWAP32(u)
3749
#endif
3850

3951
typedef ZEND_SET_ALIGNED(1, uint16_t unaligned_uint16_t);
@@ -42,41 +54,23 @@ typedef ZEND_SET_ALIGNED(1, uint64_t unaligned_uint64_t);
4254
typedef ZEND_SET_ALIGNED(1, unsigned int unaligned_uint);
4355
typedef ZEND_SET_ALIGNED(1, int unaligned_int);
4456

45-
/* Mapping of byte from char (8bit) to long for machine endian */
46-
static int byte_map[1];
47-
48-
/* Mappings of bytes from int (machine dependent) to int for machine endian */
49-
static int int_map[sizeof(int)];
50-
51-
/* Mappings of bytes from shorts (16bit) for all endian environments */
52-
static int machine_endian_short_map[2];
53-
static int big_endian_short_map[2];
54-
static int little_endian_short_map[2];
55-
56-
/* Mappings of bytes from longs (32bit) for all endian environments */
57-
static int machine_endian_long_map[4];
58-
static int big_endian_long_map[4];
59-
static int little_endian_long_map[4];
60-
61-
#if SIZEOF_ZEND_LONG > 4
62-
/* Mappings of bytes from quads (64bit) for all endian environments */
63-
static int machine_endian_longlong_map[8];
64-
static int big_endian_longlong_map[8];
65-
static int little_endian_longlong_map[8];
66-
#endif
67-
6857
/* {{{ php_pack */
69-
static void php_pack(zval *val, size_t size, int*map, char *output)
58+
static void php_pack(constzval *val, size_t size, php_pack_endiannessendianness, char *output)
7059
{
71-
size_t i;
72-
char *v;
73-
74-
convert_to_long(val);
75-
v = (char *) &Z_LVAL_P(val);
60+
zend_ulong zl = zval_get_long(val);
7661

77-
for (i = 0; i < size; i++) {
78-
*output++ = v[map[i]];
62+
if ((endianness == PHP_LITTLE_ENDIAN) != MACHINE_LITTLE_ENDIAN) {
63+
zl = PHP_LONG_BSWAP(zl);
64+
#if MACHINE_LITTLE_ENDIAN
65+
zl >>= (sizeof(zl) - size) * 8;
66+
#endif
67+
} else {
68+
#if !MACHINE_LITTLE_ENDIAN
69+
zl <<= (sizeof(zl) - size) * 8;
70+
#endif
7971
}
72+
73+
memcpy(output, (const char *) &zl, size);
8074
}
8175
/* }}} */
8276

@@ -88,10 +82,7 @@ ZEND_ATTRIBUTE_CONST static inline uint16_t php_pack_reverse_int16(uint16_t arg)
8882
/* {{{ php_pack_reverse_int32 */
8983
ZEND_ATTRIBUTE_CONST static inline uint32_t php_pack_reverse_int32(uint32_t arg)
9084
{
91-
uint32_t result;
92-
result = ((arg & 0xFF) << 24) | ((arg & 0xFF00) << 8) | ((arg >> 8) & 0xFF00) | ((arg >> 24) & 0xFF);
93-
94-
return result;
85+
return ZEND_BYTES_SWAP32(arg);
9586
}
9687
/* }}} */
9788

@@ -509,7 +500,7 @@ PHP_FUNCTION(pack)
509500
case 'c':
510501
case 'C':
511502
while (arg-- > 0) {
512-
php_pack(&argv[currentarg++], 1, byte_map, &ZSTR_VAL(output)[outputpos]);
503+
php_pack(&argv[currentarg++], 1, PHP_MACHINE_ENDIAN, &ZSTR_VAL(output)[outputpos]);
513504
outputpos++;
514505
}
515506
break;
@@ -518,16 +509,16 @@ PHP_FUNCTION(pack)
518509
case 'S':
519510
case 'n':
520511
case 'v': {
521-
int*map = machine_endian_short_map;
512+
php_pack_endiannessendianness = PHP_MACHINE_ENDIAN;
522513

523514
if (code == 'n') {
524-
map = big_endian_short_map;
515+
endianness = PHP_BIG_ENDIAN;
525516
} else if (code == 'v') {
526-
map = little_endian_short_map;
517+
endianness = PHP_LITTLE_ENDIAN;
527518
}
528519

529520
while (arg-- > 0) {
530-
php_pack(&argv[currentarg++], 2, map, &ZSTR_VAL(output)[outputpos]);
521+
php_pack(&argv[currentarg++], 2, endianness, &ZSTR_VAL(output)[outputpos]);
531522
outputpos += 2;
532523
}
533524
break;
@@ -536,7 +527,7 @@ PHP_FUNCTION(pack)
536527
case 'i':
537528
case 'I':
538529
while (arg-- > 0) {
539-
php_pack(&argv[currentarg++], sizeof(int), int_map, &ZSTR_VAL(output)[outputpos]);
530+
php_pack(&argv[currentarg++], sizeof(int), PHP_MACHINE_ENDIAN, &ZSTR_VAL(output)[outputpos]);
540531
outputpos += sizeof(int);
541532
}
542533
break;
@@ -545,16 +536,16 @@ PHP_FUNCTION(pack)
545536
case 'L':
546537
case 'N':
547538
case 'V': {
548-
int*map = machine_endian_long_map;
539+
php_pack_endiannessendianness = PHP_MACHINE_ENDIAN;
549540

550541
if (code == 'N') {
551-
map = big_endian_long_map;
542+
endianness = PHP_BIG_ENDIAN;
552543
} else if (code == 'V') {
553-
map = little_endian_long_map;
544+
endianness = PHP_LITTLE_ENDIAN;
554545
}
555546

556547
while (arg-- > 0) {
557-
php_pack(&argv[currentarg++], 4, map, &ZSTR_VAL(output)[outputpos]);
548+
php_pack(&argv[currentarg++], 4, endianness, &ZSTR_VAL(output)[outputpos]);
558549
outputpos += 4;
559550
}
560551
break;
@@ -565,16 +556,16 @@ PHP_FUNCTION(pack)
565556
case 'Q':
566557
case 'J':
567558
case 'P': {
568-
int*map = machine_endian_longlong_map;
559+
php_pack_endiannessendianness = PHP_MACHINE_ENDIAN;
569560

570561
if (code == 'J') {
571-
map = big_endian_longlong_map;
562+
endianness = PHP_BIG_ENDIAN;
572563
} else if (code == 'P') {
573-
map = little_endian_longlong_map;
564+
endianness = PHP_LITTLE_ENDIAN;
574565
}
575566

576567
while (arg-- > 0) {
577-
php_pack(&argv[currentarg++], 8, map, &ZSTR_VAL(output)[outputpos]);
568+
php_pack(&argv[currentarg++], 8, endianness, &ZSTR_VAL(output)[outputpos]);
578569
outputpos += 8;
579570
}
580571
break;
@@ -1178,127 +1169,3 @@ PHP_FUNCTION(unpack)
11781169
}
11791170
}
11801171
/* }}} */
1181-
1182-
/* {{{ PHP_MINIT_FUNCTION */
1183-
PHP_MINIT_FUNCTION(pack)
1184-
{
1185-
int i;
1186-
1187-
if (MACHINE_LITTLE_ENDIAN) {
1188-
/* Where to get lo to hi bytes from */
1189-
byte_map[0] = 0;
1190-
1191-
for (i = 0; i < (int)sizeof(int); i++) {
1192-
int_map[i] = i;
1193-
}
1194-
1195-
machine_endian_short_map[0] = 0;
1196-
machine_endian_short_map[1] = 1;
1197-
big_endian_short_map[0] = 1;
1198-
big_endian_short_map[1] = 0;
1199-
little_endian_short_map[0] = 0;
1200-
little_endian_short_map[1] = 1;
1201-
1202-
machine_endian_long_map[0] = 0;
1203-
machine_endian_long_map[1] = 1;
1204-
machine_endian_long_map[2] = 2;
1205-
machine_endian_long_map[3] = 3;
1206-
big_endian_long_map[0] = 3;
1207-
big_endian_long_map[1] = 2;
1208-
big_endian_long_map[2] = 1;
1209-
big_endian_long_map[3] = 0;
1210-
little_endian_long_map[0] = 0;
1211-
little_endian_long_map[1] = 1;
1212-
little_endian_long_map[2] = 2;
1213-
little_endian_long_map[3] = 3;
1214-
1215-
#if SIZEOF_ZEND_LONG > 4
1216-
machine_endian_longlong_map[0] = 0;
1217-
machine_endian_longlong_map[1] = 1;
1218-
machine_endian_longlong_map[2] = 2;
1219-
machine_endian_longlong_map[3] = 3;
1220-
machine_endian_longlong_map[4] = 4;
1221-
machine_endian_longlong_map[5] = 5;
1222-
machine_endian_longlong_map[6] = 6;
1223-
machine_endian_longlong_map[7] = 7;
1224-
big_endian_longlong_map[0] = 7;
1225-
big_endian_longlong_map[1] = 6;
1226-
big_endian_longlong_map[2] = 5;
1227-
big_endian_longlong_map[3] = 4;
1228-
big_endian_longlong_map[4] = 3;
1229-
big_endian_longlong_map[5] = 2;
1230-
big_endian_longlong_map[6] = 1;
1231-
big_endian_longlong_map[7] = 0;
1232-
little_endian_longlong_map[0] = 0;
1233-
little_endian_longlong_map[1] = 1;
1234-
little_endian_longlong_map[2] = 2;
1235-
little_endian_longlong_map[3] = 3;
1236-
little_endian_longlong_map[4] = 4;
1237-
little_endian_longlong_map[5] = 5;
1238-
little_endian_longlong_map[6] = 6;
1239-
little_endian_longlong_map[7] = 7;
1240-
#endif
1241-
}
1242-
else {
1243-
zval val;
1244-
int size = sizeof(Z_LVAL(val));
1245-
Z_LVAL(val)=0; /*silence a warning*/
1246-
1247-
/* Where to get hi to lo bytes from */
1248-
byte_map[0] = size - 1;
1249-
1250-
for (i = 0; i < (int)sizeof(int); i++) {
1251-
int_map[i] = size - (sizeof(int) - i);
1252-
}
1253-
1254-
machine_endian_short_map[0] = size - 2;
1255-
machine_endian_short_map[1] = size - 1;
1256-
big_endian_short_map[0] = size - 2;
1257-
big_endian_short_map[1] = size - 1;
1258-
little_endian_short_map[0] = size - 1;
1259-
little_endian_short_map[1] = size - 2;
1260-
1261-
machine_endian_long_map[0] = size - 4;
1262-
machine_endian_long_map[1] = size - 3;
1263-
machine_endian_long_map[2] = size - 2;
1264-
machine_endian_long_map[3] = size - 1;
1265-
big_endian_long_map[0] = size - 4;
1266-
big_endian_long_map[1] = size - 3;
1267-
big_endian_long_map[2] = size - 2;
1268-
big_endian_long_map[3] = size - 1;
1269-
little_endian_long_map[0] = size - 1;
1270-
little_endian_long_map[1] = size - 2;
1271-
little_endian_long_map[2] = size - 3;
1272-
little_endian_long_map[3] = size - 4;
1273-
1274-
#if SIZEOF_ZEND_LONG > 4
1275-
machine_endian_longlong_map[0] = size - 8;
1276-
machine_endian_longlong_map[1] = size - 7;
1277-
machine_endian_longlong_map[2] = size - 6;
1278-
machine_endian_longlong_map[3] = size - 5;
1279-
machine_endian_longlong_map[4] = size - 4;
1280-
machine_endian_longlong_map[5] = size - 3;
1281-
machine_endian_longlong_map[6] = size - 2;
1282-
machine_endian_longlong_map[7] = size - 1;
1283-
big_endian_longlong_map[0] = size - 8;
1284-
big_endian_longlong_map[1] = size - 7;
1285-
big_endian_longlong_map[2] = size - 6;
1286-
big_endian_longlong_map[3] = size - 5;
1287-
big_endian_longlong_map[4] = size - 4;
1288-
big_endian_longlong_map[5] = size - 3;
1289-
big_endian_longlong_map[6] = size - 2;
1290-
big_endian_longlong_map[7] = size - 1;
1291-
little_endian_longlong_map[0] = size - 1;
1292-
little_endian_longlong_map[1] = size - 2;
1293-
little_endian_longlong_map[2] = size - 3;
1294-
little_endian_longlong_map[3] = size - 4;
1295-
little_endian_longlong_map[4] = size - 5;
1296-
little_endian_longlong_map[5] = size - 6;
1297-
little_endian_longlong_map[6] = size - 7;
1298-
little_endian_longlong_map[7] = size - 8;
1299-
#endif
1300-
}
1301-
1302-
return SUCCESS;
1303-
}
1304-
/* }}} */

‎ext/standard/pack.h‎

Lines changed: 0 additions & 22 deletions
This file was deleted.

‎ext/standard/php_standard.h‎

Lines changed: 0 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -29,7 +29,6 @@
2929
#include "php_ext_syslog.h"
3030
#include "php_filestat.h"
3131
#include "php_browscap.h"
32-
#include "pack.h"
3332
#include "url.h"
3433
#include "pageinfo.h"
3534
#include "fsock.h"

0 commit comments

Comments
(0)

AltStyle によって変換されたページ (->オリジナル) /