Josh Sherman
Have you ever asked yourself, “I wonder which of these blocks of code will run
faster?” I know I have and I use a very basic template for setting up these
benchmarking experiments. The gist of the script is to run through each block
of code n times tracking the amount of time and memory it took to run. It’s
not nearly as complete as it could be, but it serves as a good enough starting
point for doing some quick tests.
<?php
$n = 1000000;
$start_time = microtime(true);
$start_mem = memory_get_usage(true);
for ($i = 0; $i < $n; $i++)
{
}
$time = microtime(true) - $start_time;
$mem = memory_get_usage(true) - $start_mem;
echo "Test 1\t{$time}\t{$mem}\n";
You can duplicate the test block as many times as you need it. Here’s a working
benchmark I wrote to see if PHP’s built-in array_keys function was actually
faster than doing a foreach:
<?php
$n = 1000000;
$array = [
'one' => 'This is the first',
'two' => 'This is the second',
'three' => 'This is the third',
'four' => 'This is the fourth',
'five' => 'This is the fifth',
];
$start_time = microtime(true);
$start_mem = memory_get_usage(true);
for ($i = 0; $i < $n; $i++)
{
unset($array_keys);
$array_keys = array_keys($array);
}
$time = microtime(true) - $start_time;
$mem = memory_get_usage(true) - $start_mem;
echo "array_keys()\t{$time}\t{$mem}\n";
$start_time = microtime(true);
$start_mem = memory_get_usage(true);
for ($i = 0; $i < $n; $i++)
{
unset($array_keys);
foreach ($array as $key => $value)
{
$array_keys[] = $key;
}
}
$time = microtime(true) - $start_time;
$mem = memory_get_usage(true) - $start_mem;
echo "foreach()\t{$time}\t{$mem}\n";
One improvement would be to track the average time per iteration as well as the fastest and slowest iteration. I went ahead and threw the code in a GitHub gist in case anyone wants to fork and improve upon it.
