Re: [need for advices] fastCaching
| From: | MARTY Fabien | Date: | Mon, 22 Jul 2002 13:33:41 +0000 |
| Subject: | Re: [need for advices] fastCaching | ||
| References: | 1 2 | Groups: | php.pear.dev |
| Request: | Send a blank email to pear-dev+get-7846@lists.php.net to get a copy of this message | ||
(Tomas)
> I guess that just isolating the container isCached() functions into a
> light layer would multiply the speed factor.
Ok, I'm just finishing benchs. I use this minimalist layer for tests.
Cache/Container/file.php :
<?php
class Cache_Container_file {
function isCached() {return true;}
function isExpired() {return false;}
function load() {return '123';}
}
?>
Note that the cache doesn't work because the load() method return always
'123'. Of course, before benching with this minimalist layer, I made a
'one hour file cache' with the correct version.
Results :
---------
- fastCache : 150 requests / second
- Cache : 30 requests / second
- Cache with minimalist layer : 55 requests / second
(and the Cache doestn't work, there is no file read)
So I think that this is not a solution. Maybe it's possible to lightly
improve perfs by making two layers (one very light for mosts cases where
there is a cache file, and another when there is no cache). But
certainly not do a big perfs jump without changing the conception.
And it would be a mistake to change completly PEAR/Cache because we
would loose its advantages : flexibility, more powerfull... So why not a
new cache class perfs oriented ? They don't exclude themselves.
Thanks,
Fabien