Framework Performance according to Rasmus

Alright, so invoking Rasmus in the title is a bit provocative, but I stumbled on an interesting talk of his; showing performance benchmarks for a number of popular php frameworks. The first portion of the talk is exactly what he presented @ OSCON, but the second half looks to be raw performance numbers. It looks like there are a couple specific, but easy, performance tweaks that he granted to certain frameworks, and I like seeing the data so much I thought I could try my hand at distilling it into Google Docs charts.

<img style="display:block; margin:0px auto 10px; text-align:center;cursor:pointer; cursor:hand;width: 320px; height: 227px;" src="http://4.bp.blogspot.com/_Fa_U5q7fBBY/SQC4W_KKKEI/AAAAAAAAABI/YXgaNBS3iOY/s320/framework_performance_(higher_is_better).png" border="0" alt=""id="BLOGGER_PHOTO_ID_5260407069819414594" />

<img style="display:block; margin:0px auto 10px; text-align:center;cursor:pointer; cursor:hand;width: 320px; height: 227px;" src="http://3.bp.blogspot.com/_Fa_U5q7fBBY/SQC4jP7k-cI/AAAAAAAAABQ/BN53Z--Vp30/s320/framework_performance_(lower_is_better).png" border="0" alt=""id="BLOGGER_PHOTO_ID_5260407280480090562" />

It's interesting that some tulsaphp guys and were recently talking about Zend had to be the slowest of all the frameworks, but apparently not so! That honor goes to CakePHP?

Question or comment about this post? Tell me on GitHub.

Framework Performance according to Rasmus / groovecoder by groovecoder is licensed under a Creative Commons Attribution-ShareAlike CC BY-SA
  1. Ya CakePHP is slow as an elderly woman with a bad leg crossing a 6 lane crosswalk, but dev time is faster and with proper caching it can handle the load. Wish they would do more to make it faster, but tis the life.
  2. Nice charts - that's kinda funny that so many talks were about CakePHP and it's actually the slowest by far. I've never really liked it personally, but it wasn't really for performance reasons. It was mainly the fact that it still supports PHP4 and has some rather sloppy internal code. Good followup post to OSCON. :)
Framework Performance according to Rasmus / groovecoder by groovecoder is licensed under a Creative Commons Attribution-ShareAlike CC BY-SA