comparison of caching plugins

I used to use WP Super Cache. There’re two good reasons to start with that one: the author is also a WP contributor and he can be found on these forums frequently to answer your questions.
At some point the plugin started to interfere with my dropdown menues and we couldn’t get our finger behind it, so I just stopped using it. It is not like I have that many visitors anyway…
No recent experience with other cache plugins.

nobody else?
anyone who udnerstands the difference between the caching techniques used by these plugins and can enlighten us?

Ovidiu,

Although I am relatively new to the web/php/wp scene – I have much experience in the area of performance tuning servers and midrange systems.

Ultimately, whatever one does to tweak performance one would hope to see a faster experience for the user. Caching is one of several areas that if setup properly can result in faster response times.

I have searched all over the place for information on caching techniques as they relate to php, wordpress and specific caching plugins. No luck!

My feeling is that the variety of combinations on how a WP site can be setup probably makes it difficult to come up with a general set of caching guidelines. Based on one’s own setup – one would need to trial and error (recording results along the way) as to what works best for one’s own site.

Shared hosting naturally complicates the situation as conditions on the server can change from moment to moment. From the investigations I have done so far – implementing caching on WP does not necessarily result in faster page loads times. My sense is that caching will only benefit those sites that have very high traffic.

For those sites (eg mine) that only get few hundred visitors per day I believe that greater performance benefits will be experienced by reducing CPU overhead and this can really only be achieved by having a minimal plugin footprint. WP Tuner is great tool for getting an idea of which plugins are both DB and CPU heavy.

Cheers,

R

PS. if you do find something please share

hey rschilt,

thanks for the extensive answer.

here is my situation:

I am on my own root server, MYSQL is set up pretty good, caching on its own most of the queries if possible. No need optimizing MYSQl.
So basically havin eaccelerator in place (which adds another boost but only for the vhosts running on fastCGI and not those on suphp) I think my best bet is not a caching plugin that caches queries but rather one that caches to static html files like wp super cache. But even those static files are then served by apache, so the process consumes as much memory as without caching, except that the generation time should be less.

I hope that is right, at least that is how I undestood things so far.

dunno if this will keep the formatting, but I’ll give it a try. This is an analysis of my current site. Most of the time is consumed in the start part… unfortunately I am not good enough to debug it further and find out whats taking 700ms…

Index 	Marker 	%Time 	%DB Time 	%DB Count 	Time 	DB Time 	DB Count 	Memory 	Output
 	 	 	 	 	(msec) 	(msec) 	 	(kb) 	lev(bytes)
0	Start	59	25	10	784.2	5.6	7	?
1	plugins_loaded	5	0	0	64.9	0.0	0	30,665.6	1(0)
2	widgets_init	3	3	7	39.4	0.6	5	32,482.1	1(0)
3	init	6	0	0	78.5	0.0	0	33,835.7	1(0)
4	posts_selection	14	11	8	179.2	2.6	6	35,658.5	1(0)
5	wp_head	2	2	4	23.3	0.4	3	35,860.4	1(9875)
6	loop_start	1	11	4	18.9	2.4	3	36,027.5	1(14132)
7	get_footer	7	22	52	97.4	5.0	37	36,105.1	1(17567)
8	posts_selection	0	2	6	2.9	0.5	4	36,840.3	1(42435)
9	loop_start	2	25	8	31.7	5.7	6	36,849.8	1(42449)
10	Stop	 	 	 	 	 	 	36,897.7	1(48291)
	Total	100	100	100	1,320.4	22.8	71

I asked a good friend of mine who I did my degree with, who is pretty knowledgeable about plugins, about caching worpdress and he suggested wp-super cache over the regular wp-cache.

He also mentioned that caching can cause it’s own set of problems.

So, there is some completely un-researched, un-solicited feedback and most likely un-helpful. It gives you something to read though ??

@nickscoper I think your friend is spot on when he mentioned: “caching can cause it’s own set of problems”. Every blog is different – every host is slightly different – no two dynamics exactly the same. In some cases caching will help – others… caching will add load and in other situations again – even have a negative impact on performance. Caching is certainly not the cure all.

Our friend Ovidiu is coming up with some precise measurements on his own blog – what he will discover will be great help help to him but to the rest of us merley general guidelines. I think what some of us are currently looking for are some clear,precise and technical guidelines around when a blog might benefit from caching and how to reduce CPU usage. So far none of us have found anything (and no one has come forward yet) with this kind of info.

When I think back to my performance tuning days with IBM Midrange – we used to say that tuning was both a science and an art. There’s no end to the “arty farty” side but fortunately the “science” element allowed us to come with a clear list of steps to follow which allowed us to speed up performance in 80% of cases.

To really get a handle on things and ultimately improve performance one must do this in a controlled way. Try something – measure – try something else and measure – record and compare results. Persist. Look and understand.

@ovidiu looks like ya been busy man – with wp-tuner – and you appear to have the measuring / recording and comparing under control. WP Tuner will reveal obvious performance bottlenecks but to drill down like you are wanting to do with “Start” will require further “drilling”. From what I understand “Start” is a measurement of WP basically cranking up. Every time someone visits any page on your blog – WP cranks up. It’s not like – you turn on system – let it boot up and processes are sitting there waiting for incoming requests. If you want to get a better idea of the breakup of the .7 secs being consumed by “Start” you are going to have to insert some WP_tuner hooks around the place. When you find out where – to get a handle on “Start” – please let us know.

R

Additional findings:

– Turning Caching off all together on my site has resulted in zero CPU throttling and in fact slightly less CPU utilisation overall.

– I have come to the conclusion that a WP site will only benefit from caching if the number of visitors is above a certain limit. In my case I only get a few hundred visitors per day and am now (almost) convinced that I will have a better performing site if caching is turned off altogether.

– The question then arises: How many visitors per day (or per minute/hour etc…) justifies the use of caching in a WP environment. Once again, difficult to come up with a general rule of thumb – my sense is that one’s site would need to experience many instances of a high number of concurrent visitors throughout a 24 hour period to justify implementing caching in WP.

R

check out this guy’s modified version: https://murmatrons.armadillo.homeip.net/features/experimental-eaccelerator-wp-super-cache and read the comments. there are some really good pieces of advice in there ??