Performance comparison between simple `Enum.map` and `Task.async_stream` in Elixir (oversimplified case)

bench_async_vs_map.exs

Mix.install([
  {:benchee, "~> 1.0"}
])

defmodule Cruncher do
  def async(n) do
    1..n
    |> Task.async_stream(&(&1 ** 2))
    |> Enum.reduce(0, fn {:ok, s}, acc -> s + acc end)
  end

  def map(n) do
    1..n
    |> Enum.map(&Integer.pow(&1, 2))
    |> Enum.sum()
  end
end

Benchee.run(%{
  "async" => fn -> Cruncher.async(1000) end,
  "map" => fn -> Cruncher.map(1000) end
})

# Example run:
#
# Operating System: macOS
# CPU Information: Intel(R) Core(TM) i7-8850H CPU @ 2.60GHz
# Number of Available Cores: 12
# Available memory: 32 GB
# Elixir 1.14.1
# Erlang 25.1.2
#
# Benchmark suite executing with the following configuration:
# warmup: 2 s
# time: 5 s
# memory time: 0 ns
# reduction time: 0 ns
# parallel: 1
# inputs: none specified
# Estimated total run time: 14 s
#
# Benchmarking async ...
# Benchmarking map ...
#
# Name            ips        average  deviation         median         99th %
# map         38.01 K      0.0263 ms    ±31.32%      0.0248 ms      0.0504 ms
# async       0.161 K        6.21 ms     ±7.40%        6.10 ms        8.25 ms
#
# Comparison:
# map         38.01 K
# async       0.161 K - 235.99x slower +6.18 ms