Re: [SMARTY] yet another benchmark

From: Date: Mon, 10 Jun 2002 21:30:12 +0000
Subject: Re: [SMARTY] yet another benchmark
References: 1  Groups: php.smarty.general 
Request: Send a blank email to php-smarty-general+get-4752@lists.php.net to get a copy of this message
Brian E. Lozier wrote:
Nice work. On a side note, I really see no reason to try to make Smarty fast without an accelerator. An accelerator not only speeds up the loading of the template classes, but it also speeds up every template since each one is a PHP script. And since php-acclerator is free anyways, why would you ever go without one? ;-)
I tend to agree, however most of my clients are on Windows machines. The ones that aren't are all on virtual hosting accounts, so I don't really have control over what software gets installed on the server. Now on one hand, this GENERALLY means that since they're on a shared account, they don't get that much traffic anyway... good point but a lot of ISPs are now monitering server load and if you go up past a certain percentage for xx amount of time, they automatically disable your account.
You would think that the hosting providers would be the first to put in an accelerator. It would make the response times much snappier, reduce CPU load and allow them to fit many more clients per server.
Brian
Brian E. Lozier wrote:
I loaded the benchmark on my local FreeBSD development machine and I posted the results online (they're not
in
real time). I ran them about 10 times each and the template systems never "changed places" between
running,
although the milliseconds did vary a tiny bit (never more than 3 or 4). The important difference between
these
and the ones on the smarty site is that mine DO NOT have anything like php-accelerator installed, so it's a vanilla php/apache module (the typical set up, I think). Monte: Yes, these are more fair and I'm pleased with the results. I'm glad the include()s now happen and
get
calculated to the results. These benchmarks don't really show anything different than the original
benchmarks
here: http://www.phpinsider.com/benchmarks/ Smarty is fastest with an accel since the large class doesn't have to be interpreted each time. Under
maximum
iterations, the speed of smarty shows because you're executing native PHP code, whereas with the rest,
you're
executing something like a regular expression or str_replace() on each and every iteration. For non-accel machines, it takes quite a while to make up the difference (parsing a large class and executing native code
vs
parsing a small class and calling a func or regex each iteration). Now, obviously none of these take into account the REST of the features the various templating systems
offer (or
don't offer). On this, I think we can all agree that smarty is by far at the top of the list feature-wise
(and
will likely stay that way thanks to the plugins). Brian
-- Monte Ohrt <monte@ispi.net> Director of Technology, ispi Inc. -- Smarty General Mailing List (http://smarty.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
-- Monte Ohrt <monte@ispi.net> Director of Technology, ispi Inc.

« previous php.smarty.general (#4752) next »