压力测试

压测脚本中创建了一个 PyDict ，分别读写 PHP 代码和 Python 代码执行 1000万次。

• PHP 版本：PHP 8.2.3 (cli) (built: Mar 17 2023 15:06:57) (NTS)
• Python 版本：Python 3.11.5
• 操作系统：Ubuntu 20.04
• GCC 版本：gcc version 9.4.0 (Ubuntu 9.4.0-1ubuntu1~20.04.2)

请注意此测试需要构造一个 1000 万元素的 HashTable，需要至少 2G 以上内存空间才可以运行

PHP

<?php
$dict = new PyDict();
const COUNT = 10000000;
$n = COUNT;
$s = microtime(true);
while ($n--) {
    $dict['key-' . $n] = $n * 3;
}
echo 'dict set: ' . round(microtime(true) - $s, 6) . ' seconds' . PHP_EOL;
$c = 0;
$n = COUNT;
$s = microtime(true);
while ($n--) {
    $c += $dict['key-' . $n];
}
echo 'dict get: ' . round(microtime(true) - $s, 6) . ' seconds' . PHP_EOL;

Python

import time
my_dict = {}
COUNT = 10000000
n = COUNT
start_time = time.time()
for i in range(n):
    my_dict["key-" + str(i)] = i * 3
elapsed_time = time.time() - start_time
print(f"dict set: {elapsed_time:.6f} seconds")
n = COUNT
total = 0
start_time_get = time.time()
for i in range(n):
    total += my_dict["key-" + str(i)]
elapsed_time_get = time.time() - start_time_get
print(f"dict get: {elapsed_time_get:.6f} seconds")

PHP 数组

<?php
ini_set('memory_limit', '2G');
$dict = [];
const COUNT = 10000000;
$n = COUNT;
$s = microtime(true);
while ($n--) {
    $dict['key-' . $n] = $n * 3;
}
echo 'array set: ' . round(microtime(true) - $s, 6) . ' seconds' . PHP_EOL;
$c = 0;
$n = COUNT;
$s = microtime(true);
while ($n--) {
    $c += $dict['key-' . $n];
}
echo 'array get: ' . round(microtime(true) - $s, 6) . ' seconds' . PHP_EOL;

结果对比

(base) htf@swoole-12:~/workspace/python-php/docs/benchmark$ php dict.php 
dict set: 4.663758 seconds
dict get: 3.980076 seconds
(base) htf@swoole-12:~/workspace/python-php/docs/benchmark$ php array.php 
array set: 1.578963 seconds
array get: 0.831129 seconds
(base) htf@swoole-12:~/workspace/python-php/docs/benchmark$ python dict.py 
dict set: 5.321664 seconds
dict get: 4.969081 seconds
(base) htf@swoole-12:~/workspace/python-php/docs/benchmark$

以 Python 测试为基准：

• phpy 以 PHP 代码写入 PyDict 的性能比原生 Python 高 14%，读取性能高 25%
• PHP 写入 PHP Array 的性能比 Python 写入 Dict 高 237%，读取高出了近 500%