String#casecmp is faster than String#downcase. We can save some CPU cycles by replacing String#downcase with String#casecmp.

The following is a script to benchmark.

#!/usr/bin/env ruby

def measure(p)
  t1 = Time.now
  1000000.times {p.call}
  t2 = Time.now
  t2 - t1
end

str1 = (1..100).map {('A'..'z').to_a.sample}.join
str2 = str1.downcase

puts "String#casecmp  :  #{measure(->{ str1.casecmp(str2) == 0})} sec"
puts "String#downcase :  #{measure(->{ str1.downcase == str2 })} sec"

Here is a result. casecmp is remarkably faster than downcase.

laphroaig:patches maeda$ ruby casecmp-vs-downcase.rb
String#casecmp  :  0.358929 sec
String#downcase :  0.557218 sec