Benchmark of Hash#merge vs. Ruby 2.0 keyword arguments.

hash_merge_benchmark.rb

#!/usr/bin/env ruby

require 'benchmark'

Benchmark.bm do |b|
  n = 1_000_000

  hash1 = {a: 1, b: 2, c: 3}
  hash2 = {x: 1, y: 2, z: 3}

  b.report('Hash#merge') do
    n.times { {a: 1, b: 2, c: 3}.merge(hash2) }
  end

  b.report('keyword args') do
    n.times { {a: 1, b: 2, c: 3, **hash2} }
  end
end

       user     system      total        real
Hash#merge  1.540000   0.000000   1.540000 (  1.550766)
keyword args  0.650000   0.000000   0.650000 (  0.648758)

@postmodern the results are much closer with Hash#update:

       user     system      total        real
Hash#merge  1.750000   0.010000   1.760000 (  1.757309)
Hash#update  0.750000   0.000000   0.750000 (  0.752853)
keyword args  0.700000   0.010000   0.710000 (  0.709861)

EDIT: apparently I didn't understand the context of the benchmark: https://twitter.com/raggi/status/317903577668534272