curl https://gist.github.com/yury/5248925/raw/string_vs_bytes.rb | ruby

string_vs_bytes.rb

require 'benchmark'

def byte_can_encode(str)
  str.each_char do |c|
    # simple range checks for most common characters (' '..'_') or ('a'..'~')
    b = c.getbyte(0)
    unless b >= 0x20 && b <= 0x5F || b >= 0x61 && b <= 0x7E
      return false
    end
  end
  true
end

def string_can_encode(str)
  str.each_char do |c|
    # simple range checks for most common characters (' '..'_') or ('a'..'~')
    unless c >= ' ' && c <= '_' || c >= 'a' && c <= '~'
      return false
    end
  end
  true
end


Benchmark.bm do |r|
  N = 100_000
  POSITIVE = (' '..'_').to_a.join('')
  NEGATIVE = POSITIVE + 'привет'

  r.report("string positive") do
    N.times { string_can_encode(POSITIVE)}
  end

  r.report("byte   positive") do    
    N.times { byte_can_encode(POSITIVE) }
  end

  r.report("string negative") do
    N.times { string_can_encode(NEGATIVE)}
  end

  r.report("byte   negative") do    
    N.times { byte_can_encode(NEGATIVE) }
  end
end

#        user     system      total        real
# string positive  3.120000   0.000000   3.120000 (  3.112871)
# byte   positive  1.150000   0.000000   1.150000 (  1.156764)
# string negative  3.210000   0.000000   3.210000 (  3.202055)
# byte   negative  1.190000   0.000000   1.190000 (  1.196076)