This post originated from an RSS feed registered with Ruby Buzz by Antonio Cangiano. Original Post: Holy Shmoly, Ruby 1.9 smokes Python away! Feed Title: Zen and the Art of Ruby Programming Feed URL: http://programmingzen.com/category/ruby/feed/ Feed Description: Antonio Cangiano's blog about Ruby development. Latest Ruby Buzz Posts Latest Ruby Buzz Posts by Antonio Cangiano Latest Posts From Zen and the Art of Ruby Programming

Alright the title of this post is a tad sensational sounding, I know, and it’s in part aimed at messing with my many Pythonista friends. We Rubyists have been teased for a long time, due to the slowness of the main Ruby interpreter. Well, it looks like with Ruby 1.9, it’ll be payback time. Just out of curiosity I decided to run a single benchmark (you can hardly call it that) to see how Ruby 1.9 had improved over the current stable version (1.8.6). I wasn’t planning to make a post about it. It was one of those tests that you do at 3 AM in an irb session when you feel you’ve made your daily peace with your actual workload for the night. When I saw the results though, my jaw dropped. I had to blog about this one.

I ran a recursive Fibonacci function, just to stress test a bit of recursion and method calling, and while I was at it, I decided to compare it with Python too. The test was run on Mac OS X 10.5 with my MacBook Pro (Core 2 Duo 2.2 GHz and 2 GB of memory). It’s a single test (which is obviously not a real world example, as you would use an iterative version of the function if it were), and unlike with real programs, it doesn’t stress many features of the language. At least for now, there is no reasonable evidence to conclude that Ruby 1.9 - which will be released for this coming Christmas - will actually be faster than Python 2.5.1 in the majority of situations, but hear me out and check out these very surprising results.

The Ruby code:

def fib(n)
  if n == 0 || n == 1
    n
  else
    fib(n-1) + fib(n-2)
  end
end

36.times do |i|
  puts "n=#{i} => #{fib(i)}"
end

And the Python equivalent:

def fib(n):
   if n == 0 or n == 1:
      return n
   else:
      return fib(n-1) + fib(n-2)

for i in range(36):
    print "n=%d => %d" % (i, fib(i))

Running the snippets above, I got the following results:

Ruby 1.8.6:       158.869s
Python 2.5.1:      31.507s
Ruby 1.9.0:        11.934s

Ehm, hold on a second! Did Ruby just go from 159 seconds down to 12? Koichi Sasada, do you have an Amazon Wishlist? I was expecting a decent improvement, as I’ve been playing with 1.9 every now and then for a long time - so I knew it was faster - but I was blown away when I timed the latest version from trunk (even if it’s a really silly example that’s being tested). Granted Python is not the fastest language out there, but Ruby 1.9 was still able to execute the script almost 3 times as fast. It’s unbelievable.

Now it’ll be very interesting to run a series of algorithmically equivalent tests for Ruby and Python, and to see just when exactly Ruby 1.9 manages to knock Python out of the water - and where Python has still the edge. If I manage to find some time, I will report the results in this blog. But for now, I’ll say just… wow!