added warmed up jruby benchmarks

README.markdown

@@ -33,4 +33,5 @@ By default, only the 4 top results get retrieved, you can specify the exact amou
 ## Experimentations
 
 Here are some benchmarks, it looks like running multiple concurrent threads is often not worth it
-http://gist.github.com/45350
+http://gist.github.com/45350
+warmed up jruby benchmarks

script/jruby-benchmarks.rb

@@ -0,0 +1,24 @@
+# For better performance, you have to warm up JRuby
+
+require 'benchmark'
+require File.join(File.dirname(__FILE__), '..', 'lib', 'ruby-web-search')
+
+10.times do
+  n = 50000
+  Benchmark.bm(50) do |x|
+    x.report("search")                          { RubyWebSearch::Google.search(:query => "Natalie Portman") }
+    x.report("unthreaded_search")               { RubyWebSearch::Google.unthreaded_search(:query => 'Natalie Portman') }
+
+    x.report("search :size => 30")              { RubyWebSearch::Google.search(:query => "Natalie Portman", :size => 30) }
+    x.report("unthreaded_search :size => 30")   { RubyWebSearch::Google.unthreaded_search(:query => 'Natalie Portman', :size => 30) }
+
+    x.report("search :size => 80")              { RubyWebSearch::Google.search(:query => "Natalie Portman", :size => 80) }
+    x.report("unthreaded_search :size => 80")   { RubyWebSearch::Google.unthreaded_search(:query => 'Natalie Portman', :size => 80) }
+
+    x.report("search :size => 150")             { RubyWebSearch::Google.search(:query => "Natalie Portman", :size => 150) }
+    x.report("unthreaded_search :size => 150")  { RubyWebSearch::Google.unthreaded_search(:query => 'Natalie Portman', :size => 150) }
+
+    x.report("search :size => 200")             { RubyWebSearch::Google.search(:query => "Natalie Portman", :size => 200) }
+    x.report("unthreaded_search :size => 200")  { RubyWebSearch::Google.unthreaded_search(:query => 'Natalie Portman', :size => 200) }  
+  end
+end