If you set WP-FFPC to APC, Memcache will not be used at all by WP-FFPC.
Zend Opcache has nothing to do with WP-FFPC: what every opcache does at first is to compile PHP and store the compiled version. APC and Xcache had an additional key-value storage as well which could be used by the user; APCu is this and only this part of APC.
As you can see I cannot optimize WP-FFPC for Opcache at max for APCu which is not done more than to Memcached/Memcache.
As far as I can see, the difference is Memcache: are you using an persitent object cache plugin as well?
Hi
No, I wasn’t using a persistent object cache plugin as well. Didn’t know hey existed. Just installed one for APCu, it speeded things up a bit.
Check this then: http://wordpress.org/plugins/apcu/
It gives a bit of an extra kick 🙂
Yep, that was the one I used. Not really a plugin as it doesn’t reside in the plugins folder nor appear in the installed plugins. Whilst it isn’t practical, the best speed boost comes from deactivating WooCommerce. My theme (Ultimatum Theme), also has an overhead. One thing I did not notice was that using a minify plugin SIGNIFICANTLY slows the page load down. I tried a lot of different ones, they all had the same effect.
Don’t use minify; gzip should be more than enough.
OK. Thanks for the advice. It lowers the Page Speed and Yslow scores, but improves page load speed. FYI, I tested all the free minify plugins and Autoptimize was the best.
ah, autoptimize, I know a thing or two about that … 😉
autoptimize typically does the following;
1. aggregation of code (css and/or js)
2. returning minified code from cache that if available
3. minification of aggregated code
4. caching of aggregated & minified code
in a setup without a page cache, AO will for the first request for a page do 1-4. for the second request for a page with exactly the same code it will only do 1&2.
in a setup with a page cache (as wp-ffpc), AO will for the first request (i.e. not a hit in the page cache and not in the ao-cache) do 1-4 as well, but will, for subsequent request do … nothing as the page cache will contain a reference to the aggregated & minified code in ao-cache.
so although for “cold” requests there definitively a performance hit (main impact is from the minification), this is not the case for “warm” requests, esp. if a page cache is in place.
hope that clarifies some.
have a nice w-e,
frank
Thanks for the clarification on AO. That explains why the page tests I did show a longer first hit response with AO in place, but faster repeat hit.
