Forums / Install & configuration / Best php cache for eZ Publish?
Andreas Kaiser
Tuesday 31 March 2009 6:26:50 am
For Drupal, some time ago somebody made a comparison of APC, eAccelerator and Xcache: http://2bits.com/articles/benchmarking-drupal-with-php-op-code-caches-apc-eaccelerator-and-xcache-compared.html
eZ System recommends an accelerator like APC, eAccelerator, Zend Platform.
We are using APC, eAccelerator and Xcache in different servers (but we can't make a comparison because some servers are client ones...)
What accelerator are you using/think it's the best for ezp 4.x?
eZ Partner in Madrid (Spain) Web: http://www.atela.net/
André R.
Tuesday 31 March 2009 7:56:33 am
In PHP 4 we used to recommend eAccelerator or APC, in that order because eAccelerator had slightly better performance.
In PHP 5 however the gap is close to non existence, even seen APC performing better with eZ Publish in some benchmarks. APC also seems to be the only actively developed php op cacher, so I would personally recommend APC.
Other arguments: APC will be part of PHP 6.0 and the article you point to mentions segmentation faults in xcache and eAccelerator.
eZ Online Editor 5: http://projects.ez.no/ezoe || eZJSCore (Ajax): http://projects.ez.no/ezjscore || eZ Publish EE http://ez.no/eZPublish/eZ-Publish-Enterprise-Subscription @: http://twitter.com/andrerom
Wednesday 01 April 2009 8:26:11 am
Thanks Andre,
We also used eaccelerator with ezp3x (PHP4) with great results, because of some other benchmarks that show eaccelerator performed not very good with PHP5 we used APC and Xcache in ezp 4 (PHP5)
The funny is that in other articles I read that APC had segmentation faults (and Xcache didn't)...
So I really opened this topic to have feedback of comments about php accelerators in relation with ez publish...
But now knowing that APC will be part of PHP6 I will go only with APC - the good is that we just installed APC in our last server installation!
Gaetano Giunta
Wednesday 01 April 2009 10:58:55 am
I have no specific info about speed differences of php caches in relation to eZP, all I can say is that I have seen all (of the free ones) crash and burn on different setups, and behave extra-stable on others. It is porobably a matter of different php application, php version and compilation options, linked libraries and os...
APC is for sure the one with more traction, as there are more devs behind it, and most of them are php kernel devs. The main problem with both eAccelerator and XCache looks like the small number of developers that hack on them.
Principal Consultant International Business Member of the Community Project Board
Damien Pobel
Thursday 02 April 2009 2:03:30 am
Hi,
About a year ago I made a benchmark between eZ Publish 3.10 and eZ Publish 4 [1] with xcache, eAccelerator, APC and without any opcode cache.
With PHP4, eAccelarator was the best (about 10%), but with PHP5 there was nearly no difference. So I think the best opcode cache is probably the one you're able to configure correctly. And as André said, PHP6 should be bundled with APC, so it can also be a good choice right now.
[1] http://pwet.fr/blog/benchmark_between_ez_publish_4_and_ez_publish_3_10_with_or_without_a_php_opcode_cache
Cheers !
Damien Planet eZ Publish.fr : http://www.planet-ezpublish.fr Certification : http://auth.ez.no/certification/verify/372448 Publications about eZ Publish : http://pwet.fr/tags/keywords/weblog/ez_publish
anny sharing
Sunday 26 December 2010 11:07:27 pm
Cache reduces the server's pressure to calcute things(pretty simple definition right?) especially MySQL queries, which can become numerous on a sigle page; Mix this with a few dozen visitors on your website in a single minute who will each visit websites with particular queries and you could easily get up to hundreds of queries per minute,That could slow your server, and caching stuff makes the charge quite lighter.
If you want to know more free PHP cache script you can check out at PHPKode.com
Script start
Module start 'content'
Module end 'content'
Script end
Time used to render debug report: 0.0002 secs