I think one of the greatest contributions of the Merb guys is spreading the message that Ruby is not slow. Even Rails is faster than most mainstream PHP frameworks. Wow.
(I looked up the research they refer to to see if they don’t cheat by comparing Rails and Merb to CakePHP. They don’t, it is the fastest PHP framework from the sample group.
[UPDATE: Nope, was wrong; please read the comments.])
3 Comments
Actually, CodeIgniter http://codeigniter.com/ is the fastest PHP framework we tested. The problem is that PHP has to reload the entire framework for each request. On the other hand, in Ruby land we optimized our system and went down to C when we needed speed, which explains why Ruby for the Web is way faster than PHP for the web, even though, as a language, PHP is faster.
Once again, people should benchmark stuff they would use the language for, not stuff they won’t use (i.e: fibonacci)
OK so I was wrong in the end. (Since higher is better in the table) CakePHP was the slowest one. So why use it for comparison when CodeIgniter is still four times slower than Rails?
And isn’t it a problem of deployment/runtime of PHP frameworks then?
I’m not sure I follow you on the “Ruby on the Web is faster than PHP while the PHP language is faster than Ruby.” Isn’t it the same Ruby running “on the web”?
I should go watch the keynote I guess..
Oh web servers are optimized in C. Now that makes sense.
Really recommend everyone who hasn’t yet to watch the talk: http://www.merbcamp.com/video/katz3.mp4