[wp-hackers] wp-cache2 performance (was apache 2.2.2 upgrade)

[Angsuman Chakraborty]

I haven't explored into WP's caching mechanism. So my statement of similarity was restricted to the fact that they both (wp2 & wp cache 2) use caching. I now realize they employ pretty different methods.

My test configuration was:
Web Server machine & clients:
Pentium 4 1.8Ghz running on 256MB RAM. 
No special configuration or optimization was done other than increasing the threads on Apache server to 250.
I am using a custom theme with lots of plugins. You can see it working at http://test.taragana.net/

I am very much interested to see someone repeat the test in a better configuration and compare the results.

The difference I observed was statistically significant. I can get more data on Monday if you are interested.

Best,
Angsuman

Simple Thoughts <http://blog.taragana.com/>   Hot Computer Jobs



-----Original Message-----
From: wp-hackers-bounces at lists.automattic.com
[mailto:wp-hackers-bounces at lists.automattic.com]On Behalf Of David Chait
Sent: Saturday, July 15, 2006 9:27 PM
To: wp-hackers at lists.automattic.com
Subject: Re: [wp-hackers] wp-cache2 performance (was apache 2.2.2
upgrade)
Importance: Low


Just so we're clear (not just you and me, but people who don't know this 
stuff):  WP2 and WP-Cache are NOT 'caching the same way', but are entirely 
different caching systems (unless someone snuck WP-Cache's code into 
WP2.0.3...).

WP2 stores an 'object cache', basically the results of certain queries on 
the database.  In that sense, it's really a MySQL cache.

WP-Cache/Staticize is an output-buffering system, which captures the results 
of the dynamic page generation to a static PHP file on disk, and on 
subsequent requests sees the static data is there and just spits it back out 
and exits (before the major WP code has to load, before the DB gets hit --  
well maybe once or twice...).

Those are two very different approaches.  Again, each could be heavily 
dependent on system configuration, to the level that if a system were 
configured a certain way, and a particular type of test were run, they could 
look approximate equals.  I wouldn't be shocked if with a really fast 
machine, lots of ram, an opcode cache, and a big MySQL query cache, NEITHER 
one would show an order of magnitude improvement.

Oh, and using lighttpd (or commercial speedy-server equiv) instead of 
apache. ;)

-d

----- Original Message ----- 
Saturday, July 15, 2006 12:56 AM,  Angsuman Chakraborty wrote:
|
|> Otherwise, I'd be shocked that loading a cache file from disk (which 
should
|> be in disk cache/memory for frequent data) takes longer than processing 
the
|> main query, plus running the various transformation code, plus plugins,
|> plus...  Just wouldn't make physical sense based on the code execution.
|> Well, again, unless disk is slow and cpu is fast -- AND maybe you have an
|> opcode cache running, which would significantly impact performance.
|
|You are forgetting WordPress 2.0.3 also does caching the same way as 
wp-cache 2. So it is
|really a comparison between caching systems.

_______________________________________________
wp-hackers mailing list
wp-hackers at lists.automattic.com
http://lists.automattic.com/mailman/listinfo/wp-hackers