benchmark.md

source: ruby tapas 322 benchmarking

some examples https://gist.github.com/equivalent/3c9a4c9d07fff79062a3

require 'benchmark'

result = Benchmark.measure do
  10_000.times do 
    str = 'Supercalifragisilticexpialidocious'
    str.reverse!
    File.write('/tmp/word.txt', str)
  end
end

result.real   # the actual time that passed (wall clock time)
              # this value can be affected by other stuff the comupter
              # is doing
#=> 0.754803923

result.utime  # time spent in user land (ruby virtual machine) 
              # (not affected by other stuff comutper is doing)
#=> 0.14

result.stime  # time spent in operating kernel (file write)
              # for execution of this piece of code (not affected by
              # other stuff comutper is doing)
#=> 0.34

result.total  # utime + stime 
#=> 0.48000000000000004 

require 'benchmark'

class Handler
  def handle_stepintime
  end
end

handler = Handler.new
n       = 1_000_000

Benchmark.bmbm(allign_number_of_chars = 20) do |reporter_object|
  event = :stepintime

  reporter_object.report('case dispatch') do
    n.times do
      case event
      when :stepintime then handler.handle_stepintime
      end
    end
  end

  reporter_object.report('dynamic_dispatch') do
    n.times do
      handler.send("handle_#{event}")
    end
  end
end

#                            user     system      total        real
# case dispatch          0.120000   0.000000   0.120000 (  0.113965)
# dynamic_dispatch       0.500000   0.000000   0.500000 (  0.503731)

when virtual machine isn't wormed up it will affect the results (so second may be faster than first just because it's lunched as second)

we can preworm VM with bmbm that will lunch it twice so the second result is more accuret

Benchmark.bmbm(allign_number_of_chars = 20) do |reporter_object|
  event = :stepintime

  reporter_object.report('case dispatch') do
    n.times do
      case event
      when :stepintime then handler.handle_stepintime
      end
    end
  end

  reporter_object.report('dynamic_dispatch') do
    n.times do
      handler.send("handle_#{event}")
    end
  end
end



# Rehearsal --------------------------------------------------------
# case dispatch          0.110000   0.000000   0.110000 (  0.114776)
# dynamic_dispatch       0.500000   0.000000   0.500000 (  0.500196)
# ----------------------------------------------- total: 0.610000sec
#
#                          user     system      total        real
# case dispatch          0.120000   0.000000   0.120000 (  0.112768)
# dynamic_dispatch       0.510000   0.000000   0.510000 (  0.507379)