In [1]:

    

require 'matplotlib/iruby'
Matplotlib::IRuby.activate

require 'pycall/import'
include PyCall::Import

pyimport :pandas, as: :pd
pyimport :seaborn, as: :sns


    

    
Out[1]:





:sns

In [2]:

    

require 'benchmark'


    

    
Out[2]:





true

In [3]:

    

n = 100_000
trials = 100
array = Array.new(n) { rand }
enum = array.each
method = []
runtime = []


    

    
Out[3]:





[]

In [4]:

    

# Array#sum
trials.times do
  method << 'array.sum'
  runtime << Benchmark.realtime { array.sum }
end


    

    
Out[4]:





100

In [5]:

    

# Array#sum
trials.times do
  method << 'enum.sum'
  runtime << Benchmark.realtime { enum.sum }
end


    

    
Out[5]:





100

In [6]:

    

# Array#inject
trials.times do
  method << 'array.inject'
  runtime << Benchmark.realtime { array.inject :+ }
end


    

    
Out[6]:





100

In [7]:

    

# Enumerable#inject
trials.times do
  method << 'enum.inject'
  runtime << Benchmark.realtime { enum.inject :+ }
end


    

    
Out[7]:





100

In [8]:

    

# while
def while_sum(array)
  sum, i, cnt = 0, 0, array.length
  while i < cnt
    sum += array[i]
    i += 1
  end
  sum
end

trials.times do
  method << 'while'
  runtime << Benchmark.realtime { while_sum(array) }
end


    

    
Out[8]:





100

In [9]:

    

df = pd.DataFrame.new(data: {method: method, runtime: runtime})
df.groupby('method').describe()


    

    
Out[9]:





  

    

      

      
runtime
    
    

      

      
count
      
mean
      
std
      
min
      
25%
      
50%
      
75%
      
max
    
    

      
method
      

      

      

      

      

      

      

      

    
  
  

    

      
array.inject
      
100.0
      
0.013103
      
0.001518
      
0.012029
      
0.012350
      
0.012590
      
0.013337
      
0.022215
    
    

      
array.sum
      
100.0
      
0.001882
      
0.000372
      
0.001498
      
0.001587
      
0.001804
      
0.002008
      
0.003009
    
    

      
enum.inject
      
100.0
      
0.023528
      
0.002267
      
0.021078
      
0.021782
      
0.023041
      
0.024574
      
0.033859
    
    

      
enum.sum
      
100.0
      
0.010069
      
0.001304
      
0.009078
      
0.009221
      
0.009463
      
0.010708
      
0.018378
    
    

      
while
      
100.0
      
0.011932
      
0.001089
      
0.011079
      
0.011357
      
0.011491
      
0.012085
      
0.018309
    
  

In [10]:

    

sns.barplot(x: 'method', y: 'runtime', data: df, errwidth: 2.5, capsize: 0.04)
plt = Matplotlib::Pyplot
plt.title("Array and Enumerable summation benchmark (#{trials} trials)")
plt.xlabel("Summation method")
plt.ylabel("Average runtime [sec]")


    

    
Out[10]:












    
Out[10]:





#<Object:0x007fd93f020088 @__pyptr__=#<PyCall::PyPtr:0x007fd93f020010 type=Text addr=0x000001176abe10>>

In [ ]: