Prev: i wanna join
Next: Variable (Class instantiation) collision
From: Nathan Rixham on 8 Oct 2010 12:16
chris h wrote:
> Saeed here's a quick (and dirty) test I ran:
>
>
> $tests = 1000000;
>
> $start = microtime(true);
> for ($i=0; $i<$tests; $i++) {
>
> $a = md5( rand() );
> $b = md5( rand() );
>
> $c = $a.$b;
> }
> var_dump( "By concat op:\t". (microtime(true) - $start) );

that's not a fair test because you have rand() and md5() calls in there
(something temporally varying)

Here's a quick test script which does 100 million iterations on both, 3
times to get some half measurable results

$i = $its = 100000000;
$tests = 3;
$a = 'foo';
$b = 'bar';

while($tests-->0) {
$t = microtime(true);
while($i-->0) {
$c = "$a$b";
}
echo 'time .: ' . (microtime(true)-$t) . PHP_EOL;
$i = $its;
$t = microtime(true);
while($i-->0) {
$c = $a.$b;
}
echo 'time ": ' . (microtime(true)-$t) . PHP_EOL;
}

I also ran the tests in the opposite order just to ensure they were
fair, results are that $a.$b (concatenation) averaged 22 seconds, and
the "$a$b" approach was 28 seconds.

Thus, concatenation is faster - but you have to get up to circa 10
million+ uses per second to use it.

Best,

Nathan
First  |  Prev  | 
Pages: 1 2 3 4 5
Prev: i wanna join
Next: Variable (Class instantiation) collision