Geek Answers Handbook

Strange performance difference between PHP versions of Ubuntu and Macports

After some profiling in a personal pet project, I play with some silly micro-optimizations. This is pretty much an experiment more than anything else, so the things that I'm setting up don't really need to be set up. This is still an interesting exercise.

In any case, I came across a strange performance difference between my installation of PHP 5.3 on OS X via Macports and on Ubuntu with apt.

The following code seems to show a big speed difference between two different versions of OS X, but only a small speed difference on Ubuntu.

$x = array(9); // As per BarsMonster comment, this ensures it runs for more // than a second in order to avoid possible kernel scheduler differences $count = 10000000; $s = microtime(true); for ($i=0;$i<$count;$i++) { $q = is_string($x); } var_dump(microtime(true)-$s); $s = microtime(true); for ($i=0;$i<$count;$i++) { // This is obviously only useful if you'll never need a string // with 'Array' in it here $q = (string)$x!='Array'; } var_dump(microtime(true)-$s);

Exit when working in OS X:

 float(17.977133989334) float(4.2555270195007)

Exit when working on Ubuntu:

 float(5.2112979888916) float(3.4337821006775)

It doesn't surprise me that the numbers for the hacked version of the cast are pretty similar, but the is_string method is very different.

What can this be connected with? If performance changes so much from installation to installation for trivial type testing functions, how can I trust the results of profiling using an OS that does not match my target deployment platform?

There is no time difference when turning APC on or off on both Ubuntu and OS X.

+4
source share
2 answers

Was installing Ubuntu on the same PC as installing OS X?

It may be something related to zts.

Try to configure it yourself with the same flags on both systems. The default options for ubuntu may be enabled, which are disabled by default on ports.

I don’t know exactly what, but your phpinfo () will say. If this is not the case (maybe phpinfo doesn't show, maybe it just shows a blank page, and then maybe one of your systems compiles with the Susoshin patch (or something else).

+1
source

  • Test for 1 second minumum, for less than 0.1 seconds, differences in the kernel planner (and dozens of other reasons) can affect your results.

  • Check if any accelerators are installed by accident (APC, eAccelerator, etc) - just compare phpinfo () from both instances and you will see.

0
source

More articles:


All Articles