Rough benchmarking of running the (roughly) equivalent sorting code in ruby vs sql

rb_vs_sql.rb

require 'active_record'
require 'logger'

# This connection will do for database-independent bug reports.
ActiveRecord::Base.establish_connection 'sqlite3:///:memory:'

ActiveRecord::Schema.define do
  create_table(:posts)    { |t| t.string :name }
  create_table(:comments) { |t| t.integer :post_id }
end

class Post < ActiveRecord::Base
  has_many :comments
end

class Comment < ActiveRecord::Base
  belongs_to :post
end

10000.times do |pi|
  p = Post.create!(:name => ->{Array.new(10){[*'a'..'z'].sample }.join}[])
  3.times { Comment.create!(:post_id => p.id) }
end

require 'pry'
require 'benchmark'

binding.pry
Benchmark.bm(7) do |x|
  x.report("SQL:")  { (1..100).each { Post.find_by_sql('SELECT posts.* FROM posts ORDER BY posts.name ASC;') }}
  x.report("AREL:") { (1..100).each { Post.all.sort_by(&:name)     }}
  x.report("RUBY:") { (1..100).each { Post.all.to_a.sort_by(&:name)}}
end
              user     system      total        real
#SQL:     18.540000   0.160000  18.700000 ( 18.756530)
#AREL:    20.160000   0.180000  20.340000 ( 20.396667)
#RUBY:    21.160000   0.170000  21.330000 ( 21.441187)

While this was running I checked on my CPU

and on the ram usage