Benchmark startup and shutdown

[Azu]

Hello

 

I can benchmark the time between the startup and the shutdown, but how can I benchmark the time of the actual startup and shutdown themselves?

[Wuhtzu]

Startup and shutdown (of what) ?

[Azu]

PHP

[ToonMariner]

get your stopwatch out....

[d22552000]

LOL, not really.

 

the same way you would time a page loading, call a php script from php.

 

<?PHP

(start timer here)
require(samp.php);
(end timer here)

?>

 

<?PHP
?>

 

Now if your server doenst ignore the file altogether, this should give you the time it takes to parse and load an entire php file.  (engine times).  just devide by two and u have shutodnw and startup I guess... otherwise you can't.

 

lol, rememerb to use a 14 digit floating point number, or you might get 0 as the time.

[Azu]

Thanks, but by the time it sees the include it has already started up, and it won't shutdown until after the include and after the timers..

[d22552000]

yes but it registers another engine (php.exe) or (php-cgi.exe) per request or page load.

 

as far as my system works it seems to start a process PER include and require, os thats how I time mine..

 

I get outputs like this:

 

0.00000000000119
0.00000000000110
0.00000000000197
0.00000000000211
0.00000000000128
0.00000000000087

 

and so on.

[Azu]

Hmm.. so this will be accurate for measuring the differance in startup/shutdown times between differant extensions being used for? Because that is what I'm trying to use this for.

Like to see if mysqli adds less to startup/shutdown time, or if mysql does, for example.

[d22552000]

it should, hwy dont you try it out and tell me the output?  remember to use micro time with a 14 digit floating point number

[trq]

PHP and all its extensions are loaded when the server starts, not when your script comes across calls to a particular extension. Unless you are using dl to load extensions manually at runtime.

 

Measuring execution times this way may help indicate which extensions process certain operations quicker, but its no indication of the extension startup/shutdown times.

[trq]

Oh man. I'm having a shoker today. It would appear my description above only relates to the cgi and cli SAPI's. I'll just be quiet.

[d22552000]

lol but if your php is a module, its modules load independantly per process (itsnot smart but hey, php did it, not me)

[trq]

What exactly do you meen by module? There are no modules in php.

[Azu]

PHP and all its extensions are loaded when the server starts, not when your script comes across calls to a particular extension. Unless you are using dl to load extensions manually at runtime.

 

Measuring execution times this way may help indicate which extensions process certain operations quicker, but its no indication of the extension startup/shutdown times.

Thanks! I think maybe something is different in my PHP installation though.

If, say, I move one of the extension DLLs away, then on PHP startup, PHP gives an error and crashes.

If I start up my server without the DLL there, it starts fine. I put it back before opening a PHP page, and it works fine.

So I'm pretty sure that the extensions are loaded by PHP on every page load for me for some reason..

[trq]

Sorry, read my last reply. You can get more detail about the startup process here.

[Azu]

Oh okay thanks.

 

Well if anyone knows of like a third party program or something that can benchmark how long it takes for PHP to startup I'd really appreciate that ^^

 

I think maybe PHP can benchmark itself on this since it can't do anything until it has started up.. and can't measure itself as it is shutting down..

[d22552000]

I eman if php is copiled as a module to the web server.  as verse to the alternatice, cgi.

[Azu]

I'm running PHP as CGI.