Ruby Benchmark: dynamic anonymous class definition vs. named class

benchmark.rb

require 'benchmark'

N = 500_000

Benchmark.bm(15) do |x|
  x.report("named class") do
    class A
      def initialize
        @a = 1234
      end
      def run
        @a
      end
    end

    N.times{ A.new.run }
  end

  x.report("anonymous class") do
    N.times{
      Class.new{
        def initialize
          @a = 1234
        end
        def run
          @a
        end
      }.new.run
    }
  end
end

result.txt

                      user     system      total        real
named class       0.250000   0.010000   0.260000 (  0.247367)
anonymous class   5.230000   0.170000   5.400000 (  5.403844)

Not a surprising result, but this shows that anyone using lots of anonymous classes should consider if possible to use named/defined ones.