My previous post about using Silex and AppServer-in-PHP similarly to ExpressJS generate quite a bit of interest. In the Hacker News thread there was a question about memory usage, and so I put the AppServer under siege. Memory usage stayed constant at:

13958 bergie    20   0  125m  10m 2136 S    1  0.6   0:00.28 php        
13959 bergie    20   0  125m  10m 2136 S    1  0.6   0:00.31 php 

Then I did the same Siege for the ExpressJS version:

14051 bergie    20   0  615m  12m 4988 S    0  0.7   0:00.34 node

The PHP implementation served 3.80 requests per second, while the Node.js version served 3.79, both on 11" MacBook Air with Ubuntu 11.04.

This trivial example isn't obviously the big test of PHP app serving vs. Node.js, but at least I was surprised that PHP did better than Node here.

Source: http://bergie.iki.fi/blog/php_can_perform_better_than_node-js/