maps_bench.ex

defmodule MyMap do
  def bodyrecursive([]) do [] end
  def bodyrecursive([h|t]) do [h + 1 | bodyrecursive(t)] end

  def generic_bodyrecursive([], _) do [] end
  def generic_bodyrecursive([h|t], fun) do [fun.(h) | bodyrecursive(t)] end

  def tail(list) do tail(list, []) end
  defp tail([], acc) do :lists.reverse(acc) end
  defp tail([h|t], acc) do tail(t, [h + 1 | acc]) end

  def generic_tail(list, fun) do generic_tail(list, fun, []) end
  defp generic_tail([], _, acc) do :lists.reverse(acc) end
  defp generic_tail([h|t], fun, acc) do generic_tail(t, fun, [fun.(h) | acc]) end


  def map_tco(list, function) do Enum.reverse _map_tco([], list, function) end
  defp _map_tco(acc, [head | tail], function) do _map_tco([function.(head) | acc], tail, function) end
  defp _map_tco(acc, [], _function) do acc end

  def map_tco_concat(acc \\ [], list, function)
  def map_tco_concat(acc, [head | tail], function) do map_tco_concat(acc ++ [function.(head)], tail, function) end
  def map_tco_concat(acc, [], _function) do acc end

  def map_tco_no_reverse(list, function) do _map_tco([], list, function) end
end



map_fun = fn(i) -> i + 1 end
inputs = %{
  "1-Small 1 Thousand"    => Enum.to_list(1..1_000),
  "2-Middle 100 Thousand" => Enum.to_list(1..100_000),
  "3-Big 10 Million"      => Enum.to_list(1..10_000_000)
}

Benchee.run(%{
  "tail" =>
    fn list -> list |> MyMap.tail end,

  "generic_tail fn" =>
    fn list -> list |> MyMap.generic_tail(fn x -> x + 1 end) end,

  "generic_tail &" =>
    fn list -> list |> MyMap.generic_tail(&(&1 + 1)) end,

  "bodyrecursive" =>
    fn list -> list |> MyMap.bodyrecursive end,

  "generic_bodyrecursive fn" =>
    fn list -> list |> MyMap.generic_bodyrecursive(fn x -> x + 1 end) end,

  "generic_bodyrecursive &" =>
    fn list -> list |> MyMap.generic_bodyrecursive(&(&1 + 1)) end,

  "stdlib fn" =>
    fn list -> list |> Enum.map(fn x -> x + 1 end) end,

  "stdlib &" =>
    fn list -> list |> Enum.map(&(&1 + 1)) end,

  "list comprehension" =>
    fn list -> for x <- list do x + 1 end end,

  "map with tail and reverse" =>
    fn(list) -> MyMap.map_tco(list, map_fun) end,

  "map tco no reverse" =>
    fn(list) -> MyMap.map_tco_no_reverse(list, map_fun) end,

  },
  time: 15,
  warmup: 5,
  inputs: inputs,
  formatters: [
    &Benchee.Formatters.HTML.output/1,
    &Benchee.Formatters.Console.output/1
  ],
  html: [file: "report.html"]
)

# Erlang/OTP 19 [erts-8.2] [source] [64-bit] [smp:4:4] [async-threads:10] [hipe] [kernel-poll:false] [dtrace]
# Elixir 1.4.1
# Benchmark suite executing with the following configuration:
# warmup: 5.0s
# time: 15.0s
# parallel: 1
# inputs: 1-Small 1 Thousand, 2-Middle 100 Thousand, 3-Big 10 Million
# Estimated total run time: 660.0s


# Benchmarking with input 1-Small 1 Thousand:
# Benchmarking bodyrecursive...
# Benchmarking generic_bodyrecursive &...
# Benchmarking generic_bodyrecursive fn...
# Benchmarking generic_tail &...
# Benchmarking generic_tail fn...
# Benchmarking list comprehension...
# Benchmarking map tco no reverse...
# Benchmarking map with tail and reverse...
# Benchmarking stdlib &...
# Benchmarking stdlib fn...
# Benchmarking tail...

# Benchmarking with input 2-Middle 100 Thousand:
# Benchmarking bodyrecursive...
# Benchmarking generic_bodyrecursive &...
# Benchmarking generic_bodyrecursive fn...
# Benchmarking generic_tail &...
# Benchmarking generic_tail fn...
# Benchmarking list comprehension...
# Benchmarking map tco no reverse...
# Benchmarking map with tail and reverse...
# Benchmarking stdlib &...
# Benchmarking stdlib fn...
# Benchmarking tail...

# Benchmarking with input 3-Big 10 Million:
# Benchmarking bodyrecursive...
# Benchmarking generic_bodyrecursive &...
# Benchmarking generic_bodyrecursive fn...
# Benchmarking generic_tail &...
# Benchmarking generic_tail fn...
# Benchmarking list comprehension...
# Benchmarking map tco no reverse...
# Benchmarking map with tail and reverse...
# Benchmarking stdlib &...
# Benchmarking stdlib fn...
# Benchmarking tail...
# Generated report_1-small_1_thousand.html
# Generated report_2-middle_100_thousand.html
# Generated report_3-big_10_million.html

# ##### With input 1-Small 1 Thousand #####
# Name                                ips        average  deviation         median
# tail                            63.07 K       15.85 μs   ±195.72%       13.00 μs
# bodyrecursive                   58.57 K       17.07 μs   ±143.66%       15.00 μs
# generic_bodyrecursive &         56.47 K       17.71 μs   ±121.19%       16.00 μs
# generic_bodyrecursive fn        55.32 K       18.08 μs   ±119.32%       16.00 μs
# stdlib fn                       30.56 K       32.73 μs   ±128.55%       30.00 μs
# map tco no reverse              30.49 K       32.80 μs    ±61.16%       31.00 μs
# stdlib &                        28.34 K       35.28 μs    ±72.55%       32.00 μs
# generic_tail fn                 27.21 K       36.75 μs   ±167.29%       32.00 μs
# list comprehension              26.55 K       37.67 μs   ±394.51%       31.00 μs
# generic_tail &                  26.48 K       37.76 μs    ±64.60%       33.00 μs
# map with tail and reverse       25.80 K       38.75 μs   ±145.98%       34.00 μs

# Comparison: 
# tail                            63.07 K
# bodyrecursive                   58.57 K - 1.08x slower
# generic_bodyrecursive &         56.47 K - 1.12x slower
# generic_bodyrecursive fn        55.32 K - 1.14x slower
# stdlib fn                       30.56 K - 2.06x slower
# map tco no reverse              30.49 K - 2.07x slower
# stdlib &                        28.34 K - 2.23x slower
# generic_tail fn                 27.21 K - 2.32x slower
# list comprehension              26.55 K - 2.38x slower
# generic_tail &                  26.48 K - 2.38x slower
# map with tail and reverse       25.80 K - 2.44x slower

# ##### With input 2-Middle 100 Thousand #####
# Name                                ips        average  deviation         median
# generic_bodyrecursive &          519.02        1.93 ms    ±20.57%        1.79 ms
# generic_bodyrecursive fn         499.03        2.00 ms    ±46.20%        1.89 ms
# bodyrecursive                    494.41        2.02 ms    ±21.15%        1.91 ms
# tail                             428.99        2.33 ms    ±40.93%        2.13 ms
# stdlib &                         261.78        3.82 ms    ±19.01%        3.63 ms
# map tco no reverse               257.84        3.88 ms    ±30.36%        3.65 ms
# stdlib fn                        245.02        4.08 ms    ±24.77%        3.83 ms
# generic_tail &                   237.28        4.21 ms    ±30.86%        3.92 ms
# list comprehension               234.36        4.27 ms    ±28.82%        4.00 ms
# generic_tail fn                  215.86        4.63 ms    ±37.15%        4.29 ms
# map with tail and reverse        197.28        5.07 ms    ±40.31%        4.73 ms

# Comparison: 
# generic_bodyrecursive &          519.02
# generic_bodyrecursive fn         499.03 - 1.04x slower
# bodyrecursive                    494.41 - 1.05x slower
# tail                             428.99 - 1.21x slower
# stdlib &                         261.78 - 1.98x slower
# map tco no reverse               257.84 - 2.01x slower
# stdlib fn                        245.02 - 2.12x slower
# generic_tail &                   237.28 - 2.19x slower
# list comprehension               234.36 - 2.21x slower
# generic_tail fn                  215.86 - 2.40x slower
# map with tail and reverse        197.28 - 2.63x slower

# ##### With input 3-Big 10 Million #####
# Name                                ips        average  deviation         median
# map tco no reverse                 2.51      399.11 ms    ±19.71%      373.50 ms
# tail                               1.84      543.93 ms    ±41.75%      486.18 ms
# list comprehension                 1.41      707.87 ms    ±19.18%      675.98 ms
# map with tail and reverse          1.40      714.90 ms    ±18.26%      686.56 ms
# generic_tail fn                    1.40      714.99 ms    ±21.67%      666.39 ms
# generic_tail &                     1.20      832.02 ms    ±24.42%      803.23 ms
# generic_bodyrecursive fn           0.98     1015.60 ms    ±22.63%     1056.96 ms
# generic_bodyrecursive &            0.90     1114.73 ms    ±24.02%     1130.33 ms
# stdlib &                           0.88     1142.06 ms    ±19.41%     1219.39 ms
# stdlib fn                          0.87     1145.33 ms    ±19.42%     1231.18 ms
# bodyrecursive                      0.77     1292.77 ms    ±23.20%     1342.22 ms

# Comparison: 
# map tco no reverse                 2.51
# tail                               1.84 - 1.36x slower
# list comprehension                 1.41 - 1.77x slower
# map with tail and reverse          1.40 - 1.79x slower
# generic_tail fn                    1.40 - 1.79x slower
# generic_tail &                     1.20 - 2.08x slower
# generic_bodyrecursive fn           0.98 - 2.54x slower
# generic_bodyrecursive &            0.90 - 2.79x slower
# stdlib &                           0.88 - 2.86x slower
# stdlib fn                          0.87 - 2.87x slower
# bodyrecursive                      0.77 - 3.24x slower



# ------------------------------------------------------------------------------



# Erlang/OTP 19 [erts-8.2] [source] [64-bit] [smp:4:4] [async-threads:10] [hipe] [kernel-poll:false] [dtrace]
# Elixir 1.4.1
# Benchmark suite executing with the following configuration:
# warmup: 2.0s
# time: 5.0s
# parallel: 1
# inputs: 1-Small 1 Thousand, 2-Middle 100 Thousand, 3-Big 10 Million
# Estimated total run time: 231.0s


# Benchmarking with input 1-Small 1 Thousand:
# Benchmarking bodyrecursive...
# Benchmarking generic_bodyrecursive &...
# Benchmarking generic_bodyrecursive fn...
# Benchmarking generic_tail &...
# Benchmarking generic_tail fn...
# Benchmarking list comprehension...
# Benchmarking map tco no reverse...
# Benchmarking map with tail and reverse...
# Benchmarking stdlib &...
# Benchmarking stdlib fn...
# Benchmarking tail...

# Benchmarking with input 2-Middle 100 Thousand:
# Benchmarking bodyrecursive...
# Benchmarking generic_bodyrecursive &...
# Benchmarking generic_bodyrecursive fn...
# Benchmarking generic_tail &...
# Benchmarking generic_tail fn...
# Benchmarking list comprehension...
# Benchmarking map tco no reverse...
# Benchmarking map with tail and reverse...
# Benchmarking stdlib &...
# Benchmarking stdlib fn...
# Benchmarking tail...

# Benchmarking with input 3-Big 10 Million:
# Benchmarking bodyrecursive...
# Benchmarking generic_bodyrecursive &...
# Benchmarking generic_bodyrecursive fn...
# Benchmarking generic_tail &...
# Benchmarking generic_tail fn...
# Benchmarking list comprehension...
# Benchmarking map tco no reverse...
# Benchmarking map with tail and reverse...
# Benchmarking stdlib &...
# Benchmarking stdlib fn...
# Benchmarking tail...
# Generated report_1_-_small_-_1_thousand.html
# Generated report_2_-_middle_-_100_thousand.html
# Generated report_3_-_big_-_10_million.html

# ##### With input 1-Small 1 Thousand #####
# Name                                ips        average  deviation         median
# tail                            75.77 K       13.20 μs   ±192.88%       12.00 μs
# generic_bodyrecursive &         59.03 K       16.94 μs    ±86.31%       15.00 μs
# bodyrecursive                   58.86 K       16.99 μs    ±52.39%       15.00 μs
# generic_bodyrecursive fn        55.99 K       17.86 μs   ±131.17%       15.00 μs
# stdlib fn                       31.56 K       31.69 μs   ±176.10%       29.00 μs
# stdlib &                        30.89 K       32.37 μs    ±66.49%       29.00 μs
# list comprehension              30.83 K       32.44 μs   ±160.79%       31.00 μs
# generic_tail &                  28.86 K       34.65 μs    ±64.41%       33.00 μs
# map tco no reverse              28.77 K       34.75 μs    ±49.77%       33.00 μs
# generic_tail fn                 28.55 K       35.03 μs    ±57.88%       33.00 μs
# map with tail and reverse       26.55 K       37.67 μs    ±57.48%       36.00 μs

# Comparison:
# tail                            75.77 K
# generic_bodyrecursive &         59.03 K - 1.28x slower
# bodyrecursive                   58.86 K - 1.29x slower
# generic_bodyrecursive fn        55.99 K - 1.35x slower
# stdlib fn                       31.56 K - 2.40x slower
# stdlib &                        30.89 K - 2.45x slower
# list comprehension              30.83 K - 2.46x slower
# generic_tail &                  28.86 K - 2.63x slower
# map tco no reverse              28.77 K - 2.63x slower
# generic_tail fn                 28.55 K - 2.65x slower
# map with tail and reverse       26.55 K - 2.85x slower

# ##### With input 2-Middle 100 Thousand #####
# Name                                ips        average  deviation         median
# generic_bodyrecursive &          570.31        1.75 ms    ±15.47%        1.70 ms
# bodyrecursive                    568.73        1.76 ms    ±14.63%        1.70 ms
# generic_bodyrecursive fn         568.31        1.76 ms    ±15.54%        1.70 ms
# tail                             500.11        2.00 ms    ±37.82%        1.94 ms
# stdlib &                         309.28        3.23 ms    ±13.17%        3.07 ms
# stdlib fn                        304.57        3.28 ms    ±13.54%        3.13 ms
# map tco no reverse               300.33        3.33 ms    ±13.53%        3.17 ms
# list comprehension               273.91        3.65 ms    ±24.43%        3.45 ms
# generic_tail &                   259.20        3.86 ms    ±22.43%        3.65 ms
# generic_tail fn                  258.28        3.87 ms    ±22.23%        3.68 ms
# map with tail and reverse        241.47        4.14 ms    ±21.75%        3.91 ms

# Comparison:
# generic_bodyrecursive &          570.31
# bodyrecursive                    568.73 - 1.00x slower
# generic_bodyrecursive fn         568.31 - 1.00x slower
# tail                             500.11 - 1.14x slower
# stdlib &                         309.28 - 1.84x slower
# stdlib fn                        304.57 - 1.87x slower
# map tco no reverse               300.33 - 1.90x slower
# list comprehension               273.91 - 2.08x slower
# generic_tail &                   259.20 - 2.20x slower
# generic_tail fn                  258.28 - 2.21x slower
# map with tail and reverse        241.47 - 2.36x slower

# ##### With input 3-Big 10 Million #####
# Name                                ips        average  deviation         median
# map tco no reverse                 2.14      466.34 ms    ±20.14%      437.61 ms
# tail                               2.13      468.99 ms    ±38.22%      396.29 ms
# list comprehension                 1.56      639.71 ms    ±31.52%      644.18 ms
# map with tail and reverse          1.42      703.56 ms    ±30.40%      700.50 ms
# generic_tail fn                    1.29      777.66 ms    ±27.21%      714.83 ms
# generic_tail &                     1.15      872.69 ms    ±21.90%      923.20 ms
# generic_bodyrecursive &            1.07      933.02 ms    ±33.35%     1066.12 ms
# bodyrecursive                      1.04      960.28 ms    ±31.39%     1051.07 ms
# generic_bodyrecursive fn           1.04      960.88 ms    ±32.64%     1090.40 ms
# stdlib fn                          1.03      970.59 ms    ±30.63%     1176.81 ms
# stdlib &                           1.00      998.24 ms    ±33.43%     1187.14 ms

# Comparison:
# map tco no reverse                 2.14
# tail                               2.13 - 1.01x slower
# list comprehension                 1.56 - 1.37x slower
# map with tail and reverse          1.42 - 1.51x slower
# generic_tail fn                    1.29 - 1.67x slower
# generic_tail &                     1.15 - 1.87x slower
# generic_bodyrecursive &            1.07 - 2.00x slower
# bodyrecursive                      1.04 - 2.06x slower
# generic_bodyrecursive fn           1.04 - 2.06x slower
# stdlib fn                          1.03 - 2.08x slower
# stdlib &                           1.00 - 2.14x slower