循环sizeof($ x)与循环($ x)的实际速度/基准?

时间:2017-02-28 17:18:26

标签: php performance function loops sizeof

我在一个循环中对sizeof($ x)vs $ x的速度感到困惑。这个网站:phpbench.com声称没有pre calc -count()的sizeof($ x)循环比 count()慢<%> THOUSANDS 百分比。所以我做了如下测试,但我不确定这是否是测试它的正确方法。结果表明,每个函数的时间几乎相同。所以我不明白它与phpbench网站有多大不同。

总之,如果函数a(sizeof($ unset))实际上比函数b慢得多(使用预先计算的$ unset值),我想知道一个明确的答案。我认为现在值/函数存储在内存中,因此sizeof($ x)实际上比几年前的服务器更快?

<?php

$v=0;

function a()
{
//sizeof
$unset=array('a','b','c','d','e','f','g','h','i','j','k','l','m','n','o');

for($i=0;$i<sizeof($unset);$i++) {$v=$v+1;}

return;
}

function b()
{
//pre calculated
$unset=array('a','b','c','d','e','f','g','h','i','j','k','l','m','n','o');

for($i=0;$i<15;$i++) {$v=$v+1;}

return;
}

function benchmark($func)
{

    $start = microtime(true);
    for ($i = 0; $i < 500000; $i++) {
        $func();
    }
    $end = microtime(true);
    $time = $end - $start;

    echo $func . " time: " . sprintf('%.4f', $time) . PHP_EOL.'<br>';
}

benchmark('a'); // sizeof
benchmark('b'); // count

?>

1 个答案:

答案 0 :(得分:1)

根据我的基准sizeof()count()执行相同的操作,但是为了从循环中挤出更多性能,您可以执行以下操作:

for($i=0, $c=sizeof($unset);$i<$c;$i++){
  $v=$v+1;
}

这应该会提高性能,因为它不必在每个循环周期中使用函数进行求值。

我使用的系统是here,代码如下:

$b = new \Benchmark(10000);

$b->register('sizeof', function(){
  $unset=array('a','b','c','d','e','f','g','h','i','j','k','l','m','n','o');
  for($i=0;$i<sizeof($unset);$i++) {$v=$v+1;}
});

$b->register('count', function(){
  $unset=array('a','b','c','d','e','f','g','h','i','j','k','l','m','n','o');
  for($i=0;$i<count($unset);$i++) {$v=$v+1;}
});

$b->register('count 2', function(){
  $unset=array('a','b','c','d','e','f','g','h','i','j','k','l','m','n','o');
  for($i=0, $c=count($unset);$i<$c;$i++) {$v=$v+1;}
});

$b->register('preset', function(){
  $unset=array('a','b','c','d','e','f','g','h','i','j','k','l','m','n','o');
  for($i=0;$i<15;$i++) {$v=$v+1;}
});

print_r($b->start());

结果如下:

Array
(
    [stats] => Array
        (
            [phpversion] => 7.1.1
            [itterations] => 40000
            [duration] => 0.1180682182
            [fastest] => preset
            [slowest] => sizeof
        )

    [results] => Array
        (
            [0] => Array
                (
                    [name] => preset
                    [time] => 0.0265829563
                    [average] => 0.0000026583
                    [speed] => 19.73%
                )

            [1] => Array
                (
                    [name] => count 2
                    [time] => 0.0271441936
                    [average] => 0.0000027144
                    [speed] => 18.04%
                )

            [2] => Array
                (
                    [name] => count
                    [time] => 0.0312242508
                    [average] => 0.0000031224
                    [speed] => 5.71%
                )

            [3] => Array
                (
                    [name] => sizeof
                    [time] => 0.0331168175
                    [average] => 0.0000033117
                    [speed] => 0.00%
                )

        )

)