On Thursday 18 August 2005 3:17 pm, Josh Charles wrote:

> It's bothersome because I already dedicated a project to working in
> Rails before I found this out, so I can only hope this is a Graphical
> Toolkit Issue and not a language issue.

Ruby performance with web tasks has been mentioned a great deal.  Heck, I 
think I've mentioned it in a praising way a couple times already in the past 
week.

Rails has demonstrated adequate performance for real world tasks on many 
occasions, now.  43things.com runs on Rails and they field quite a bit of 
highly dynamic traffic without any problem on a very reasonable set of 
hardware.

And while I don't use Rails, I have used Ruby exclusively for web work for 
over three years, and I get what I deem to be very good performance from it.  
Comparing web performance is very difficult because there are so many 
variables that can influence the numbers, but an app that I have in 
production for a fortune 500 company has been benchmarked by me to deliver a 
9k, modestly dynamic page in .0035 seconds on a single processor, midrange 
Linux box.  For large groups of consecutive requests of this page (not 
cached), it does 260-290 requests per second, consistently, depending on what 
other load I have on the machine at the same time.  In actual real, heavy 
usage with real data, many of the pages will be larger or more complex, and 
thus slower to deliver, but that one cheap box will still handle more traffic 
per second than this company will _ever_ throw at it, and if it did start to 
get bogged down, adding a second server at the level of performance that Ruby 
gives me is a reasonable, economical solution.

At least when it comes to web applications, Ruby is plenty fast, and Rails in 
real usage seems to deliver enough performance to handle pretty heavy usage.


Kirk Haines