RSS
 

Archive for August 20th, 2009

Setting up an environment for jRuby on Rails

20 Aug

My apologize in advance for having used a robot translator, a small help to save a little time:

Over the next few articles i will use jruby 1.3.1 with Rails 2.3.3 and this is an opportunity to write two lines about how to configure the system.
By the way, the system used is a virtualized windows xp, java sdk 6, Glassfish as application server and javadb.

Download and install the java sdk, at the moment we are at version 6 update 14.

There is also the sdk with NetBeans, an excellent IDE to manage projects ruby on rails. It could be an opportunity to try it. The package also includes the application server Glassfish V3.

Now we download and install jRuby, atm we are at version 1.3.1.

Installation is simple: unpack the zip in a path like C:rubyjruby-131

Let’s set the environment variables, right click on “My Computer” icon on the desktop -> Properties -> select the tab “Advanced” -> click on “environment variables”.

  • Add to the PATH variable the path to the bin folder where we installed jruby, for example C:rubyjruby-131bin.
  • Set the environment variable JAVA_HOME. In the “system variables” -> “New” -> as the name “JAVA_HOME” value as the installation path dell’sdk for example C:ProgrammiJavajdk1.6.0_14

It is not necessary to restart the operating system, we test the result:

C:>jruby -v
jruby 1.3.1 (ruby 1.8.6p287) (2009-06-15 2fd6c3d) (Java HotSpot(TM) Client VM 1.6.0_14) [x86-java]

Eureka! Now we’re ready to use jruby, jirb and gem. If in your system there is also classic ruby you may specify jruby-S to recall gem otherwise you can also put jruby path in the last position inside environment variable. Ok, now we can install rails 2.3.3:

C:>jruby -S gem install rails
Successfully installed activesupport-2.3.3
Successfully installed activerecord-2.3.3
Successfully installed actionpack-2.3.3
Successfully installed actionmailer-2.3.3
Successfully installed activeresource-2.3.3
Successfully installed rails-2.3.3
6 gems installed
Installing ri documentation for activesupport-2.3.3...
Installing ri documentation for activerecord-2.3.3...
Installing ri documentation for actionpack-2.3.3...
Installing ri documentation for actionmailer-2.3.3...
Installing ri documentation for activeresource-2.3.3...
Installing ri documentation for rails-2.3.3...
Installing RDoc documentation for activesupport-2.3.3...
Installing RDoc documentation for activerecord-2.3.3...
Installing RDoc documentation for actionpack-2.3.3...
Installing RDoc documentation for actionmailer-2.3.3...
Installing RDoc documentation for activeresource-2.3.3...
Installing RDoc documentation for rails-2.3.3...

We install the jdbc adapter to connect to Java DB (i don’t use jruby-S anymore, in my system is not necessary):

C:>gem install activerecord-jdbcderby-adapter
Successfully installed activerecord-jdbc-adapter-0.9.1
Successfully installed jdbc-derby-10.4.2.0
Successfully installed activerecord-jdbcderby-adapter-0.9.1
3 gems installed
Installing ri documentation for activerecord-jdbc-adapter-0.9.1...
Installing ri documentation for jdbc-derby-10.4.2.0...
Installing ri documentation for activerecord-jdbcderby-adapter-0.9.1...
Installing RDoc documentation for activerecord-jdbc-adapter-0.9.1...
Installing RDoc documentation for jdbc-derby-10.4.2.0...
Installing RDoc documentation for activerecord-jdbcderby-adapter-0.9.1...

Then install the gem for Glassfish:

C:>gem install glassfish
Successfully installed rack-1.0.0
Successfully installed glassfish-0.9.5-universal-java
2 gems installed
Installing ri documentation for rack-1.0.0...
Installing ri documentation for glassfish-0.9.5-universal-java...
Installing RDoc documentation for rack-1.0.0...
Installing RDoc documentation for glassfish-0.9.5-universal-java...

We are ready, we create a new project:

C:>rails ProvaArticolo

… in the database’s configuration file we use:

# config/database.yml
 
# JavaDB Setup
#
# You may need to copy derby.jar into
#  TODO: location C:rubyjruby-131lib
# With Java SE 6 and later this is not necessary.
development:
  adapter: derby
  database: db/development.db
 
# Warning: The database defined as 'test' will be erased and
# re-generated from your development database when you run 'rake'.
# Do not set this db to the same as development or production.
test:
  adapter: derby
  database: db/test.db
 
# Warning: The database defined as 'test' will be erased and
# re-generated from your development database when you run 'rake'.
# Do not set this db to the same as development or production.
production:
  adapter: derby
  database: db/production.db
 
# Warning: The database defined as 'test' will be erased and
# re-generated from your development database when you run 'rake'.
# Do not set this db to the same as development or production.

Let us start the server, from the prompt inside our application’s directory:

C:ProvaArticolo>glassfish
Starting GlassFish server at: 127.0.0.1:3000 in development environment...
Writing log messages to: C:/ProvaArticolo/log/development.log.
Press Ctrl+C to stop.

At http://localhost:3000/ should see the welcome page.
Now stop the server with Ctrl + C db and create the environment in development:

C:ProvaArticolo>rake db:create
(in C:/ProvaArticolo)
db/development.db already exists

We create something within the database, we start with the source of rails:

C:ProvaArticolo>jruby script/generate scaffold article name:string body:text
create app/models/
exists app/controllers/
exists app/helpers/
exists app/views/articles
create app/views/layouts/
create test/functional/
create test/unit/
create test/unit/helpers/
create public/stylesheets/
create app/views/articles/index.html.erb
create app/views/articles/show.html.erb
create app/views/articles/new.html.erb
create app/views/articles/edit.html.erb
create app/views/layouts/articles.html.erb
create public/stylesheets/scaffold.css
create app/controllers/articles_controller.rb
create test/functional/articles_controller_test.rb
create app/helpers/articles_helper.rb
create test/unit/helpers/articles_helper_test.rb
route map.resources :articles
dependency model
exists app/models/
exists test/unit/
create test/fixtures/
create app/models/article.rb
create test/unit/article_test.rb
create test/fixtures/articles.yml
exists db/migrate
create db/migrate/20090729171105_create_articles.rb

and create the table:

C:ProvaArticolo>rake db:migrate
(in C:/ProvaArticolo)
== CreateArticles: migrating =================================================
-- create_table(:articles)
-> 0.0700s
-> 0 rows
== CreateArticles: migrated (0.0700s) ========================================

Finally remove the index.html file from inside the public folder and create the initial route:

#configroutes.rb
map.root :controller => "articles"

Reboot the server and at http://localhost:3000/ this time we should see a list of articles.

If I don’t have forgotten something, jruby on rails is ready, enjoy!

 
1 Comment

Posted in Ruby

 

Add an operation to a resource in a RESTful system

20 Aug

My apologize in advance for having used a robot translator, a small help to save a little time:

To create a RESTful application with rails is very simple:

C:>rails MyApp

In a REST system, the application consists of Web resources with a set of linked transactions. With the scaffold generate a resource and the four basic functions to manage it, called CRUD (Create, Read, Update, Delete)

C:>script/generate scaffold assets name:string

We created the base from which to test this short article.

Suppose you want to add a new operation to our resource Asset, for example we want to handle the copy. We start by creating a new address by changing the scaffold created by:

#config/routes.rb
map.resources :assets, :member => { :copy => :get }

Let’s check the results:
C:>rake routes

…ensuring that there is the new path:

copy_asset GET /assets/:id/copy(.:format) {:controller=>"assets", :action=>"copy"}

The copy is useful in many contexts because it allows for the inclusion of a new resource from an existing.
Translate into code, and let’s create the new controller:

#app/controllers/assets_controller.rb
 
# GET /assets/copy
# GET /assets/copy.xml
def copy
  @asset = asset.new
  @asset.attributes = Asset.find(params[:id]).attributes
 
  respond_to do |format|
    format.html { render :new }
    format.xml  { render :xml => @asset }
  end
end

You can also create the view, depends on the transaction that has to be managed. This example uses the view of the new as they are very similar and both the recall: create.

Create or not create a new action on the resource is pure philosophy, leaving aside the question how we create a new target, we refer to a collection of resources:

#config/routes.rb
map.resources :assets, :member => { :copy => :get }, :collection => { :createcp => :put }

I used a collection because in this case does not need a reference to an existing resource, if I had used: the member for the path generation dell’uri would have required.

if we test again the routes should add this:

createcp_asset PUT /assets/createcp(.:format) {:controller=>"assets", :action=>"createcp"}

then we create the new view and form with the uri of the new action and specify the method (for detail put but I indicated to use an equivalent post):

#app/views/assets/copy.html.erb
 
<% form_for(@asset, :url => createcp_assets_path, :method=>:put do |f| %>
  <%= f.error_messages %>
  <p>
    <%= f.label t(:name) %><br />
    <%= f.text_field :name %>
  </p>
  <p>
    <%= f.submit t(:create) %>
  </p>
<% end %>

and the new action in the controller:

#app/controllers/assets_controller.rb
 
# POST /assets/createcp
# POST /assets/createcp.xml
def createcp
  @asset = Asset.new(params[:asset])
 
  #do something
 
  respond_to do |format|
    if @asset.save
      flash[:notice] = I18n.t(:created, :model => I18n.t('model.asset'))
      format.html { redirect_to assets_path }
      format.xml  { render :xml => @asset, :status => :created, :location => @asset }
    else
      format.html { render :action => "copy" }
      format.xml  { render :xml => @asset.errors, :status => :unprocessable_entity }
    end
  end
end

To invoke the new operation by the index:

#app/views/assets/index.html.erb
<%= link_to t(:copy), copy_asset_path(asset) %>

Finally, these are the words used in the example for internationalization (if you need it):

create: "Crea"
copy: "Copia"
created: "The item {{model}} was created successfully."
name: "Name"
model:
  asset: "Asset"
 
Comments Off

Posted in Ruby on Rails

 

Best practise with four ruby’s interpreters

20 Aug

It cannot be denied that versatility is one of its features, a true multi-platform. There are interpreters for many operating systems and in addition to the classic C version we have JRuby, a viable alternative for the java world and we also have Ironruby for the Microsoft’s .NET, arrived in recent days at a fairly mature version.
Not bad! Think about the benefits that a company may have: a language easy to learn but very powerful, has an excellent framework (one in all, Rails) to quickly develop applications for intranets and in addition to all this, the chance to use the java’s or .net’s libraries.

In this article we will read about these four interpreters:

  1. Ruby 1.8.6 patch 368 updated at 2009-03-31
  2. Ruby 1.9.1 patch 129 updated at 2009-05-12 revision 23412
  3. jRuby 1.3.1 (ruby 1.8.6p287) 2009-06-15 Client VM 1.6.0 update 14
  4. IronRuby 0.9.0.0 on .NET 2.0.0.0

The Purpose: To discover the main differences in performance of the individual interpreter. There are methods and techniques that lead to the same result, without such tests can be difficult to determine which is the best practice.
In tests that we will see, I have focused on the processing speed and only roughly the amount of memory used.

The system: The physical machine is a Windows XP Professional SP3 32bit with an Intel E7300 dual core with 3Mb of cache, and 3.25 Gb of ram.

physical machine

physical machine

The tests running on a virtualized system with MS Virtual PC 2007, Windows XP Pro, Java SDK 6:14,. NET Framework 3.5 SP1, single-core processor, 768MB memory. Only the heaviest test (hash) was repeated with 2Gb.

Benchmarks are divided into four categories:

  • Strings
  • Numbers
  • Array
  • Hash

These tests that I created, are limited to perform very simple operations. At the bottom there is a total, but should not be construed as an index to the actual interpreter’s performance because it changes a lot according to the tests selected and those that I used, is not said that are conditions encountered in actual use.

Strings

This is the first script:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93

require "benchmark"
include Benchmark
 
Benchmark.bm(21, "--- Total:") do|b|
  puts "Concat 1.000.000:"
  n=1_000_000
  t1 = b.report("+") do
    n.times { "a "+"b "+"c "+"d "+"e "; "f "+"g" }
  end
  t1 += b.report("<<") do
    n.times { "a "<<"b "<<"c "<<"d "<<"e "; "f "<<"g" }
  end
  t1 += b.report("#{}") do
    sa="a"; sb="b"; sc="c"; sd="d"; se="e"
    n.times { "#{sa} #{sb} #{sc} #{sd} #{se}"; "#{sb} #{sc}" }
  end
 
  puts "Add 100.000:"
  n=100_000
  t2 = b.report("+=") do
    a = ""
    n.times { a += "." }
  end
  t2 += b.report("<<") do
    a = ""
    n.times { a << "." }
  end
  t2 += b.report("a = a + '.'") do
    a = ""
    n.times { a = a + "." }
  end
  t2 += b.report("#{}") do
    a = ""
    n.times { a = "#{a}." }
  end
 
  puts "Other 100.000:"
  n=100_000
  str=""
  t3 = b.report("* 100:") do
    n.times { str = " abc def ghi rn" * 100 }
  end
  t3 += b.report("capitalize:") do
    n.times { str.capitalize }
  end
  t3 += b.report("upcase:") do
    n.times { str.upcase }
  end
  t3 += b.report("chomp:") do
    n.times { str.chomp }
  end
  t3 += b.report("include:") do
    n.times { str.include?("ghi");str.include?("qwertyuiopasdfghjkl") }
  end
  t3 += b.report("index:") do
    n.times { str.index("ghi");str.index("qwertyuiopasdfghjkl") }
  end
  t3 += b.report("sub:") do
    n.times { str.sub("ghi", "GGHHII") }
  end
  t3 += b.report("gsub:") do
    n.times { str.gsub("ghi", "GGHHII") }
  end
  t3 += b.report("[x..y]:") do
    n.times {|x| str[0..2];str[5..15];str[6..26];str[10..50] }
  end
  t3 += b.report("slice:") do
    n.times {|x| str.slice(0..2);str.slice(5..15);str.slice(6..26);str.slice(10..50) }
  end
  t3 += b.report("strip:") do
    n.times {|x| "  1 2 3 4 5  ".strip }
  end
  t3 += b.report("Each:") do
    n.times { str.each_line{|x| x } }
  end
 
  puts "Cast 1.000.000:"
  n=1_000_000
  str="abcd1234"
  t4 = b.report(".to_i:") do
    n.times { str.to_i }
  end
  t4 += b.report(".to_sym:") do
    n.times { str.to_sym }
  end
  n=100_000
  str="abc123 " * 100
  t4 += b.report("split:") do
    n.times { str.split }
  end
 
  [t1+t2+t3+t4]
end

These results are sorted by interpreter, from first to quarter:

Ruby 1.8.6

C:>ruby bench_str.rb
                           user     system      total        real
Concat 1.000.000:
+                      3.525000   0.000000   3.525000 (  3.555325)
<<                     2.864000   0.000000   2.864000 (  2.884320)
#{}                    3.565000   0.010000   3.575000 (  3.615415)
Add 100.000:
+=                     5.257000   4.306000   9.563000 (  9.724565)
<<                     0.030000   0.010000   0.040000 (  0.040060)
a = a + '.'            5.398000   4.376000   9.774000 (  9.884805)
#{}                    5.328000   2.474000   7.802000 (  7.901835)
Other 100.000:
* 100:                 0.551000   0.380000   0.931000 (  0.981470)
capitalize:            0.951000   0.361000   1.312000 (  1.311965)
upcase:                2.844000   0.490000   3.334000 (  3.365040)
chomp:                 0.491000   0.301000   0.792000 (  0.861290)
include:               0.631000   0.000000   0.631000 (  0.630945)
index:                 0.611000   0.000000   0.611000 (  0.610915)
sub:                   1.181000   0.410000   1.591000 (  1.602400)
gsub:                 19.989000   1.623000  21.612000 ( 21.842715)
[x..y]:                0.351000   0.000000   0.351000 (  0.360540)
slice:                 0.350000   0.000000   0.350000 (  0.350525)
strip:                 0.090000   0.000000   0.090000 (  0.090135)
Each:                  7.711000   0.000000   7.711000 (  7.761625)
Cast 1.000.000:
.to_i:                 0.221000   0.000000   0.221000 (  0.220330)
.to_sym:               0.240000   0.000000   0.240000 (  0.240360)
split:                 6.229000   0.010000   6.239000 (  6.269390)
--- Total:            68.408000  14.751000  83.159000 ( 84.105970)

Ruby 1.9.1

C:>ruby bench_str.rb
                           user     system      total        real
Concat 1.000.000:
+                      1.932000   0.000000   1.932000 (  1.992985)
<<                     1.633000   0.000000   1.633000 (  1.672505)
#{}                    2.183000   0.000000   2.183000 (  2.213315)
Add 100.000:
+=                     5.538000   4.136000   9.674000 (  9.884805)
<<                     0.030000   0.000000   0.030000 (  0.030045)
a = a + '.'            5.368000   4.346000   9.714000 (  9.884805)
#{}                    6.529000   4.246000  10.775000 ( 11.096620)
Other 100.000:
* 100:                 0.491000   0.400000   0.891000 (  0.891335)
capitalize:            4.406000   0.501000   4.907000 (  4.977455)
upcase:                0.671000   0.501000   1.172000 (  1.171755)
chomp:                 0.541000   0.340000   0.881000 (  0.911365)
include:               0.110000   0.000000   0.110000 (  0.110165)
index:                 0.120000   0.000000   0.120000 (  0.120180)
sub:                   0.731000   0.371000   1.102000 (  1.111665)
gsub:                  7.361000   0.991000   8.352000 (  8.512750)
[x..y]:                0.200000   0.010000   0.210000 (  0.210315)
slice:                 0.170000   0.000000   0.170000 (  0.170255)
strip:                 0.090000   0.000000   0.090000 (  0.090135)
Each:                  6.359000   0.020000   6.379000 (  6.419615)
Cast 1.000.000:
.to_i:                 0.171000   0.000000   0.171000 (  0.170255)
.to_sym:               0.290000   0.000000   0.290000 (  0.330495)
split:                 3.575000   0.000000   3.575000 (  3.595385)
--- Total:            48.499000  15.862000  64.361000 ( 65.568205)

JRuby

C:>jruby bench_str.rb
                           user     system      total        real
Concat 1.000.000:
+                      1.142000   0.000000   1.142000 (  1.101000)
<<                     0.932000   0.000000   0.932000 (  0.942000)
#{}                    0.430000   0.000000   0.430000 (  0.430000)
Add 100.000:
+=                     7.962000   0.000000   7.962000 (  7.962000)
<<                     0.020000   0.000000   0.020000 (  0.020000)
a = a + '.'            7.542000   0.000000   7.542000 (  7.542000)
#{}                   16.324000   0.000000  16.324000 ( 16.324000)
Other 100.000:
* 100:                 0.140000   0.000000   0.140000 (  0.140000)
capitalize:            0.781000   0.000000   0.781000 (  0.781000)
upcase:                1.382000   0.000000   1.382000 (  1.382000)
chomp:                 0.051000   0.000000   0.051000 (  0.051000)
include:               0.240000   0.000000   0.240000 (  0.240000)
index:                 0.230000   0.000000   0.230000 (  0.230000)
sub:                   0.381000   0.000000   0.381000 (  0.391000)
gsub:                  3.816000   0.000000   3.816000 (  3.816000)
[x..y]:                0.100000   0.000000   0.100000 (  0.100000)
slice:                 0.120000   0.000000   0.120000 (  0.120000)
strip:                 0.040000   0.000000   0.040000 (  0.040000)
Each:                  2.333000   0.000000   2.333000 (  2.333000)
Cast 1.000.000:
.to_i:                 0.470000   0.000000   0.470000 (  0.470000)
.to_sym:               0.191000   0.000000   0.191000 (  0.191000)
split:                 1.622000   0.000000   1.622000 (  1.622000)
--- Total:            46.249000   0.000000  46.249000 ( 46.227999)

IronRuby

C:>ir bench_str.rb
                           user     system      total        real
Concat 1.000.000:
+                      2.804032   0.070101   2.874133 (  2.954425)
<<                     1.101584   0.000000   1.101584 (  1.101650)
#{}                    1.331915   0.000000   1.331915 (  1.342010)
Add 100.000:
+=                    50.502619   4.836955  55.339574 ( 56.043940)
<<                     0.040058   0.000000   0.040058 (  0.040060)
a = a + '.'           50.662850   4.616638  55.279488 ( 56.013895)
#{}                   51.173584   4.666710  55.840294 ( 56.634825)
Other 100.000:
* 100:                 0.500720   0.000000   0.500720 (  0.500750)
capitalize:            2.503600   0.010014   2.513614 (  2.583870)
upcase:                6.078741   0.000000   6.078741 (  6.199285)
chomp:                 0.200288   0.000000   0.200288 (  0.200300)
include:               4.216062   0.000000   4.216062 (  4.246360)
index:                 4.135947   0.000000   4.135947 (  4.206300)
sub:                   1.462102   0.000000   1.462102 (  1.492235)
gsub:                  8.692499   0.020029   8.712528 (  8.773140)
[x..y]:                0.290418   0.000000   0.290418 (  0.290435)
slice:                 0.230331   0.010014   0.240346 (  0.260390)
strip:                 0.070101   0.000000   0.070101 (  0.070105)
Each:                 25.015971   0.050072  25.066043 ( 25.357980)
Cast 1.000.000:
.to_i:                 0.220317   0.000000   0.220317 (  0.220330)
.to_sym:               0.280403   0.000000   0.280403 (  0.280420)
split:                 4.446394   0.010014   4.456408 (  4.516765)
--- Total:           215.960536  14.290549 230.251085 (233.329469)

This is the summary:

Ruby 1.8.6
--- Total: 68.408000 14.751000 83.159000 ( 84.105970)
Ruby 1.9.1
--- Total: 48.499000 15.862000 64.361000 ( 65.568205)
Jruby
--- Total: 46.249000 0.000000 46.249000 ( 46.227999)
Ironruby
--- Total: 215.960536 14.290549 230.251085 (233.329469)

Comments: For the concatenation, in all cases prevails, the operator << instead, include the variables in strings with # (), it may be as efficient use of memory but as a performance is at the bottom. Ironruby fatigue with the strings, the gap is huge with some operators, would enough to use << to almost cancel the gap. A little effort on the string's cycle, if compared with the other three. Very good ruby 1.9.1 with performance similar to Jruby but a memory consumption of only 5Mb against 30Mb used by java. The version 1.8.6 is not so bad, he lost time with gsub (especially compared to jruby).

Numbers

The script:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79

require "benchmark"
include Benchmark
 
Benchmark.bm(21, "--- Total:") do|b|
  puts "Integer"
  puts "Arithmetic operations 10.000.000:"
  n=10_000_000
  t1 = b.report("x + 2") do
    a = 0
    n.times {|x| a = x + 2 }
  end
  t1 += b.report("x - 1") do
    n.times {|x| a = x - 1 }
  end
  t1 += b.report("x * 3") do
    n.times {|x| a = x * 3 }
  end
  t1 += b.report("x / 2") do
    n.times {|x| a = x / 2 }
  end
  t1 += b.report("x ** 2") do
    n.times {|x| a = x ** 2 }
  end
  t1 += b.report("x % 360") do
    n.times {|x| x % 360 }
  end
  t1 += b.report("Cast") do
    n.times {|x| x.to_s }
    n.times {|x| x.to_f }
  end
 
  puts "Add 10.000.000:"
  n=10_000_000
  t2 = b.report("+=") do
    a = 0
    n.times { a += 1 }
  end
  t2 += b.report("a = a + 1") do
    a = 0
    n.times { a = a + 1 }
  end
  t2 += b.report("<<") do
    a = 0
    n.times { a << 1 }
  end
  t2 += b.report(".next") do
    a = 0
    n.times { a = a.next }
  end
 
  puts "Float"
  puts "Arithmetic operations 10.000.000:"
  n=10_000_000
  t3 = b.report("x + 1.234567") do
    a = 0
    n.times {|x| a = x + 1.234567 }
  end
  t3 += b.report("x - 0.135799") do
    a = 0
    n.times {|x| a = x - 0.135799 }
  end
  t3 += b.report("x * 0.987654") do
    a = 0
    n.times {|x| a = x * 0.987654 }
  end
  t3 += b.report("x / 1.975313") do
    a = 0
    n.times {|x| a = x / 1.975313 }
  end
  t3 += b.report("x ** 1.987654") do
    a = 0
    n.times {|x| a = x ** 1.987654 }
  end
  t3 += b.report("Cast") do
    n.times {|x| x.to_s }
    n.times {|x| x.to_i }
  end
  [t1+t2+t3]
end

Ruby 1.8.6

C:>ruby bench_num.rb
                           user     system      total        real
Integer
Arithmetic operations 10.000.000:
x + 2                  0.981000   0.000000   0.981000 (  0.991485)
x - 1                  0.901000   0.000000   0.901000 (  0.911365)
x * 3                  0.962000   0.000000   0.962000 (  0.961440)
x / 2                  0.911000   0.000000   0.911000 (  0.921380)
x ** 2                 4.316000   0.000000   4.316000 (  4.366540)
x % 360                0.881000   0.000000   0.881000 (  0.961440)
Cast                   4.727000   0.020000   4.747000 (  5.217815)
Add 10.000.000:
+=                     0.952000   0.000000   0.952000 (  0.951425)
a = a + 1              0.901000   0.010000   0.911000 (  0.951425)
<<                     1.312000   0.000000   1.312000 (  1.342010)
.next                  1.312000   0.020000   1.332000 (  1.402100)
Float
Arithmetic operations 10.000.000:
x + 1.234567           2.103000   0.000000   2.103000 (  2.233345)
x - 0.135799           1.983000   0.000000   1.983000 (  2.013015)
x * 0.987654           2.042000   0.000000   2.042000 (  2.053075)
x / 1.975313           2.183000   0.000000   2.183000 (  2.243360)
x ** 1.987654          3.174000   0.000000   3.174000 (  3.184770)
Cast                   4.106000   0.000000   4.106000 (  4.146210)
--- Total:            33.747000   0.050000  33.797000 ( 34.852200)

Ruby 1.9.1

C:>ruby bench_num.rb
                           user     system      total        real
Integer
Arithmetic operations 10.000.000:
x + 2                  3.154000   0.000000   3.154000 (  3.194785)
x - 1                  3.365000   0.000000   3.365000 (  3.415115)
x * 3                  3.415000   0.000000   3.415000 (  3.495235)
x / 2                  3.475000   0.000000   3.475000 (  3.495235)
x ** 2                14.771000   0.010000  14.781000 ( 14.952395)
x % 360                2.914000   0.000000   2.914000 (  3.004500)
Cast                  11.847000   0.000000  11.847000 ( 11.987955)
Add 10.000.000:
+=                     2.544000   0.000000   2.544000 (  2.583870)
a = a + 1              2.524000   0.000000   2.524000 (  2.533795)
<<                     2.233000   0.000000   2.233000 (  2.253375)
.next                  1.963000   0.000000   1.963000 (  1.982970)
Float
Arithmetic operations 10.000.000:
x + 1.234567           3.885000   0.000000   3.885000 (  3.985970)
x - 0.135799           3.886000   0.000000   3.886000 (  3.945910)
x * 0.987654           3.895000   0.000000   3.895000 (  3.945910)
x / 1.975313          12.378000   0.010000  12.388000 ( 12.588855)
x ** 1.987654          5.057000   0.000000   5.057000 (  5.147710)
Cast                  11.106000   0.030000  11.136000 ( 11.286905)
--- Total:            92.412000   0.050000  92.462000 ( 93.800490)

JRuby

C:>jruby bench_num.rb
                           user     system      total        real
Integer
Arithmetic operations 10.000.000:
x + 2                  2.043000   0.000000   2.043000 (  1.863000)
x - 1                  1.803000   0.000000   1.803000 (  1.803000)
x * 3                  1.923000   0.000000   1.923000 (  1.923000)
x / 2                  2.153000   0.000000   2.153000 (  2.153000)
x ** 2                 5.899000   0.000000   5.899000 (  5.899000)
x % 360                1.803000   0.000000   1.803000 (  1.803000)
Cast                   4.527000   0.000000   4.527000 (  4.527000)
Add 10.000.000:
+=                     1.682000   0.000000   1.682000 (  1.682000)
a = a + 1              1.602000   0.000000   1.602000 (  1.602000)
<<                     1.302000   0.000000   1.302000 (  1.302000)
.next                  1.433000   0.000000   1.433000 (  1.433000)
Float
Arithmetic operations 10.000.000:
x + 1.234567           2.363000   0.000000   2.363000 (  2.363000)
x - 0.135799           2.404000   0.000000   2.404000 (  2.404000)
x * 0.987654           2.153000   0.000000   2.153000 (  2.153000)
x / 1.975313           4.487000   0.000000   4.487000 (  4.487000)
x ** 1.987654          5.829000   0.000000   5.829000 (  5.839000)
Cast                   4.146000   0.000000   4.146000 (  4.156000)
--- Total:            47.552000   0.000000  47.552000 ( 47.392000)

IronRuby

C:>ir bench_num.rb
                           user     system      total        real
Integer
Arithmetic operations 10.000.000:
x + 2                  3.925645   0.060086   3.985731 (  4.056075)
x - 1                  1.652376   0.000000   1.652376 (  1.662490)
x * 3                  1.912750   0.000000   1.912750 (  1.972955)
x / 2                  1.982851   0.010014   1.992866 (  2.013015)
x ** 2                 3.895602   0.010014   3.905616 (  3.975955)
x % 360                1.672405   0.010014   1.682419 (  1.692535)
Cast                   6.248986   0.010014   6.259000 (  6.389570)
Add 10.000.000:
+=                     1.672405   0.010014   1.682419 (  1.712565)
a = a + 1              1.742506   0.000000   1.742506 (  1.792685)
<<                     1.472117   0.040058   1.512174 (  1.532295)
.next                  1.672405   0.000000   1.672405 (  1.672505)
Float
Arithmetic operations 10.000.000:
x + 1.234567           1.932779   0.000000   1.932779 (  1.932895)
x - 0.135799           1.952808   0.010014   1.962822 (  2.033045)
x * 0.987654           1.992866   0.000000   1.992866 (  1.992985)
x / 1.975313           3.825501   0.030043   3.855544 (  3.925880)
x ** 1.987654          3.014334   0.000000   3.014334 (  3.034545)
Cast                   5.878453   0.000000   5.878453 (  5.928880)
--- Total:            46.446787   0.190274  46.637061 ( 47.320875)

Summary:

Ruby 1.8.6
--- Total: 33.747000 0.050000 33.797000 ( 34.852200)
Ruby 1.9.1
--- Total: 92.412000 0.050000 92.462000 ( 93.800490)
Jruby
--- Total: 47.552000 0.000000 47.552000 ( 47.392000)
Ironruby
--- Total: 46.446787 0.190274 46.637061 ( 47.320875)

Comments: Marked improvement in the version 1.9.1 which with numbers is three times faster than 1.8.6. The other two are not bad too.
16/08/2009: Only now i see that i mistakenly put shift method (<<) under adding methods.
24/08/2009: I had inverted the 1.8.6's results with those of 1.9.1 ...actually it seemed a bit strange! :-p

Array

The script:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127

require "benchmark"
include Benchmark
 
Benchmark.bm(21, "Total:") do|b|
  n=100_000
  puts "Create 100.000:"
  t1 = b.report("%w()") do
    n.times { %w(a b c d f e g h j k i l m n o p q r s t u v w y z 0 1 2 3 4 5 6 7 8 9) }
  end
  t1 += b.report("%w''") do
    n.times { %w"a b c d f e g h j k i l m n o p q r s t u v w y z 0 1 2 3 4 5 6 7 8 9" }
  end
  t1 += b.report("split") do
    n.times { "a b c d f e g h j k i l m n o p q r s t u v w y z 0 1 2 3 4 5 6 7 8 9".split(' ') }
  end
  t1 += b.report("[str]") do
    n.times { ["a","b","c","d","f","e","g","h","j","k","i","l","m","n","o","p","q","r","s","t","u","v","w","y","z","0","1","2","3","4","5","6","7","8","9"] }
  end
  t1 += b.report("[num]") do
    n.times { [0,1,2,3,4,5,6,7,8,9,0,1,2,3,4,5,6,7,8,9,0,1,2,3,4,5,6,7,8,9,0,1,2,3,4] }
  end
  t1 += b.report("Array.new(str)") do
    n.times { Array.new(35, "a") }
  end
  t1 += b.report("Array.new(num)") do
    n.times { Array.new(35, 0) }
  end
  t1 += b.report("Array.new(hash)") do
    n.times { Array.new(35, {}) }
  end
 
  n=100_000
  puts "Add 100.000:"
  t2 = b.report("+= ['.']") do
    a = []
    n.times { a += ["."] }
  end
 
  t2 += b.report("+= [0]") do
    a = []
    n.times { a += [0] }
  end
 
  t2 += b.report("<< ['.']") do
    a = []
    n.times { a << ["."] }
  end
 
  t2 += b.report("<< [0]") do
    a = []
    n.times { a << [0] }
  end
 
  t2 += b.report("a = a+['.']") do
    a = []
    n.times { a = a + ["."] }
  end
 
  t2 += b.report("a = a+[0]") do
    a = []
    n.times { a = a + [0] }
  end
 
  n=100_000
  puts "Concatenate 100.000:"
  t3 = b.report("[num]+") do
    n.times { [0,1,2,3,4]+[5,6,7,8,9]+[0,1,2,3,4,5,6,7,8,9] }
  end
  t3 += b.report("[str]+") do
    n.times { ["a","b","c","d","f"]+["e","g","h","j","k"]+["i","l","m","n","o","p","q","r","s","t"] }
  end
  t3 += b.report("[mix]+") do
    n.times { ["a","b","c","d","f"]+[5,6,7,8,9]+["i",0,"m",1,"o",2,"q",3,"s",4] }
  end
  t3 += b.report("[num]<<") do
    n.times { [0,1,2,3,4]<<[5,6,7,8,9]<<[0,1,2,3,4,5,6,7,8,9] }
  end
  t3 += b.report("[str]<<") do
    n.times { ["a","b","c","d","f"]<<["e","g","h","j","k"]<<["i","l","m","n","o","p","q","r","s","t"] }
  end
  t3 += b.report("[mix]<<") do
    n.times { ["a","b","c","d","f"]<<[5,6,7,8,9]<<["i",0,"m",1,"o",2,"q",3,"s",4] }
  end
   t3 += b.report("[num].concat") do
    n.times { [0,1,2,3,4].concat([5,6,7,8,9]).concat([0,1,2,3,4,5,6,7,8,9]) }
  end
  t3 += b.report("[str].concat") do
    n.times { ["a","b","c","d","f"].concat(["e","g","h","j","k"]).concat(["i","l","m","n","o","p","q","r","s","t"]) }
  end
  t3 += b.report("[mix].concat") do
    n.times { ["a","b","c","d","f"].concat([5,6,7,8,9]).concat(["i",0,"m",1,"o",2,"q",3,"s",4])}
  end
  t3 += b.report("union [num]") do
    n.times { [0,14,2,3,4,53,682] | [3,4,53,6,7,84,9] }
  end
  t3 += b.report("union [str]") do
    n.times { ["a","b","c","d","f"] | ["b","d","h","j","k"] }
  end
 
  a=Array.new(100, 0)
  n=1_000_000
  puts "Read 1.000.000:"
  t4 = b.report("each:") do
    n.times { a.each{|x| x} }
  end
  t4 += b.report("map:") do
    n.times { a.map{|x| x} }
  end
 
  puts "Other 10.000:"
  n=10_000
  a=["a",nil,0,"b",1,nil] * 500
  t5 = b.report(".compact:") do
    n.times { a.compact }
  end
  t5 += b.report(".delete(nil):") do
    n.times { a.delete(nil)}
  end
 
  a=[]
  n.times {|x| a << x }
  t5 += b.report("delete:") do
    n.times {|x| a.delete(x) }
  end
  puts "-"*20
  [t1+t2+t3+t4+t5]
end

Ruby 1.8.6

C:>ruby bench_arr.rb
                           user     system      total        real
Create 100.000:
%w()                   0.300000   0.010000   0.310000 (  0.320480)
%w''                   0.300000   0.000000   0.300000 (  0.310465)
split                  2.164000   0.010000   2.174000 (  2.213315)
[str]                  0.310000   0.000000   0.310000 (  0.310465)
[num]                  0.160000   0.030000   0.190000 (  0.220330)
Array.new(str)         0.180000   0.010000   0.190000 (  0.190285)
Array.new(num)         0.151000   0.030000   0.181000 (  0.180270)
Array.new(hash)        0.270000   0.040000   0.310000 (  0.310465)
Add 100.000:
+= ['.']              22.472000   8.272000  30.744000 ( 31.487160)
+= [0]                20.800000   8.602000  29.402000 ( 30.004940)
<< ['.']               0.070000   0.010000   0.080000 (  0.090135)
<< [0]                 0.070000   0.000000   0.070000 (  0.070105)
a = a+['.']           28.321000   7.351000  35.672000 ( 36.514690)
a = a+[0]             25.226000   7.480000  32.706000 ( 33.520205)
Concatenate 100.000:
[num]+                 0.451000   0.030000   0.481000 (  0.490735)
[str]+                 0.581000   0.030000   0.611000 (  0.640960)
[mix]+                 0.521000   0.030000   0.551000 (  0.550825)
[num]<<                0.350000   0.020000   0.370000 (  0.370555)
[str]<<                0.421000   0.031000   0.452000 (  0.450675)
[mix]<<                0.420000   0.010000   0.430000 (  0.430645)
[num].concat           0.341000   0.030000   0.371000 (  0.380570)
[str].concat           0.491000   0.020000   0.511000 (  0.510765)
[mix].concat           0.400000   0.020000   0.420000 (  0.430645)
union [num]            1.032000   0.020000   1.052000 (  1.051575)
union [str]            0.761000   0.080000   0.841000 (  0.881320)
Read 1.000.000:
each:                 16.934000   0.000000  16.934000 ( 17.145680)
map:                  20.750000   1.181000  21.931000 ( 22.243315)
Other 10.000:
.compact:              0.401000   0.131000   0.532000 (  0.530795)
.delete(nil):          4.246000   0.000000   4.246000 (  4.316465)
delete:                4.937000   0.000000   4.937000 (  5.017515)
--------------------
Total:               153.831000  33.478000 187.309000 (191.186350)

Ruby 1.9.1

C:>ruby bench_arr.rb
                           user     system      total        real
Create 100.000:
%w()                   0.520000   0.000000   0.520000 (  0.520780)
%w''                   0.541000   0.000000   0.541000 (  0.550825)
split                  0.781000   0.000000   0.781000 (  0.791185)
[str]                  0.511000   0.000000   0.511000 (  0.520780)
[num]                  0.020000   0.000000   0.020000 (  0.020030)
Array.new(str)         0.170000   0.010000   0.180000 (  0.190285)
Array.new(num)         0.140000   0.000000   0.140000 (  0.140210)
Array.new(hash)        0.141000   0.030000   0.171000 (  0.200300)
Add 100.000:
+= ['.']              22.131000   9.553000  31.684000 ( 32.268330)
+= [0]                16.514000   8.422000  24.936000 ( 25.478160)
<< ['.']               0.010000   0.020000   0.030000 (  0.030045)
<< [0]                 0.030000   0.000000   0.030000 (  0.030045)
a = a+['.']           17.856000   3.946000  21.802000 ( 22.323435)
a = a+[0]             17.024000   8.172000  25.196000 ( 25.788625)
Concatenate 100.000:
[num]+                 0.211000   0.020000   0.231000 (  0.230345)
[str]+                 0.490000   0.000000   0.490000 (  0.500750)
[mix]+                 0.341000   0.000000   0.341000 (  0.340510)
[num]<<                0.140000   0.000000   0.140000 (  0.140210)
[str]<<                0.421000   0.010000   0.431000 (  0.460690)
[mix]<<                0.260000   0.000000   0.260000 (  0.260390)
[num].concat           0.160000   0.000000   0.160000 (  0.160240)
[str].concat           0.471000   0.000000   0.471000 (  0.470705)
[mix].concat           0.320000   0.000000   0.320000 (  0.320480)
union [num]            0.691000   0.020000   0.711000 (  0.721080)
union [str]            0.952000   0.000000   0.952000 (  0.961440)
Read 1.000.000:
each:                  7.861000   0.000000   7.861000 (  7.991970)
map:                  13.169000   1.482000  14.651000 ( 14.822200)
Other 10.000:
.compact:              0.250000   0.130000   0.380000 (  0.380570)
.delete(nil):          2.684000   0.000000   2.684000 (  2.704050)
delete:                4.907000   0.000000   4.907000 (  4.987470)
--------------------
Total:               109.717000  31.815000 141.532000 (144.306135)

JRuby

C:>jruby bench_arr.rb
                           user     system      total        real
Create 100.000:
%w()                   0.251000   0.000000   0.251000 (  0.211000)
%w''                   0.190000   0.000000   0.190000 (  0.190000)
split                  0.370000   0.000000   0.370000 (  0.370000)
[str]                  0.191000   0.000000   0.191000 (  0.191000)
[num]                  0.030000   0.000000   0.030000 (  0.030000)
Array.new(str)         0.080000   0.000000   0.080000 (  0.080000)
Array.new(num)         0.060000   0.000000   0.060000 (  0.060000)
Array.new(hash)        0.100000   0.000000   0.100000 (  0.100000)
Add 100.000:
+= ['.']             143.074000   0.000000 143.074000 (143.074000)
+= [0]               126.099000   0.000000 126.099000 (126.099000)
<< ['.']               0.211000   0.000000   0.211000 (  0.211000)
<< [0]                 0.090000   0.000000   0.090000 (  0.090000)
a = a+['.']          124.536000   0.000000 124.536000 (124.536000)
a = a+[0]            127.672000   0.000000 127.672000 (127.672000)
Concatenate 100.000:
[num]+                 0.120000   0.000000   0.120000 (  0.120000)
[str]+                 0.180000   0.000000   0.180000 (  0.180000)
[mix]+                 0.120000   0.000000   0.120000 (  0.120000)
[num]<<                0.070000   0.000000   0.070000 (  0.070000)
[str]<<                0.211000   0.000000   0.211000 (  0.211000)
[mix]<<                0.110000   0.000000   0.110000 (  0.110000)
[num].concat           0.100000   0.000000   0.100000 (  0.100000)
[str].concat           0.160000   0.000000   0.160000 (  0.160000)
[mix].concat           0.131000   0.000000   0.131000 (  0.131000)
union [num]            0.280000   0.000000   0.280000 (  0.280000)
union [str]            0.300000   0.000000   0.300000 (  0.300000)
Read 1.000.000:
each:                  9.364000   0.000000   9.364000 (  9.364000)
map:                  11.007000   0.000000  11.007000 ( 11.007000)
Other 10.000:
.compact:              0.411000   0.000000   0.411000 (  0.411000)
.delete(nil):          2.904000   0.000000   2.904000 (  2.904000)
delete:                4.687000   0.000000   4.687000 (  4.687000)
--------------------
Total:               553.109000   0.000000 553.109000 (553.069000)

IronRuby

C:>ir bench_arr.rb
                           user     system      total        real
Create 100.000:
%w()                   0.721037   0.000000   0.721037 (  0.721080)
%w''                   0.230331   0.000000   0.230331 (  0.250375)
split                  1.482131   0.000000   1.482131 (  1.482220)
[str]                  0.240346   0.000000   0.240346 (  0.240360)
[num]                  0.100144   0.000000   0.100144 (  0.100150)
Array.new(str)         0.150216   0.000000   0.150216 (  0.150225)
Array.new(num)         0.130187   0.000000   0.130187 (  0.130195)
Array.new(hash)        0.110158   0.000000   0.110158 (  0.120180)
Add 100.000:
+= ['.']              93.234064   9.723982 102.958046 (105.808475)
+= [0]                44.143475   3.294738  47.438213 ( 48.092030)
<< ['.']               0.160230   0.000000   0.160230 (  0.160240)
<< [0]                 0.090130   0.010014   0.100144 (  0.100150)
a = a+['.']           96.779162   9.483637 106.262798 (109.644220)
a = a+[0]             43.823014   3.545098  47.368112 ( 47.921775)
Concatenate 100.000:
[num]+                 0.160230   0.000000   0.160230 (  0.160240)
[str]+                 0.230331   0.000000   0.230331 (  0.240360)
[mix]+                 0.180259   0.000000   0.180259 (  0.180270)
[num]<<                0.090130   0.000000   0.090130 (  0.090135)
[str]<<                0.180259   0.010014   0.190274 (  0.190285)
[mix]<<                0.150216   0.000000   0.150216 (  0.150225)
[num].concat           0.110158   0.010014   0.120173 (  0.120180)
[str].concat           0.220317   0.000000   0.220317 (  0.220330)
[mix].concat           0.180259   0.000000   0.180259 (  0.180270)
union [num]            0.330475   0.000000   0.330475 (  0.340510)
union [str]            0.410590   0.000000   0.410590 (  0.410615)
Read 1.000.000:
each:                 14.040189   0.010014  14.050203 ( 14.181240)
map:                  19.848541   0.020029  19.868570 ( 20.120135)
Other 10.000:
.compact:              1.161670   0.000000   1.161670 (  1.171755)
.delete(nil):         15.462234   0.030043  15.492277 ( 15.833715)
delete:                1.422045   0.000000   1.422045 (  1.472205)
--------------------
Total:               335.572530  26.137584 361.710114 (369.984144)

Summary:

Ruby 1.8.6
Total: 153.831000 33.478000 187.309000 (191.186350)
Ruby 1.9.1
Total: 109.717000 31.815000 141.532000 (144.306135)
Jruby
Total: 553.109000 0.000000 553.109000 (553.069000)
Ironruby
Total: 335.572530 26.137584 361.710114 (369.984144)

Comments: So, in this third test is important to say a few words. Time high of jruby and ironruby is also due to the fact that I wanted to compare the concatenation += with a similar version. Both are very inappropriate for these two interpreters, times are very high and would enough to use operator << to get very different results, in order: 1.8.6 = 61 1.9.1 = 40 jruby = 33 = 60 and ironruby... jruby would be the fastest.
Another curiosity is the use of compact method, besides being more elegant is even faster than delete (nil) with sensitive differences in Ironruby.

Hash

The script to test the Hash is the heaviest as memory's stress test. Frankly, it wasn't in my intentions, but then I thought to hold it for examining the outcomes:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61

require "benchmark"
include Benchmark
 
n=1_000_000
Benchmark.bm(21, "Total:") do|b|
  puts "Write 1.000.000:"
  h = {}
  t1 = b.report("int => str:") do
    n.times {|x| h[x] = "." }
  end
  h = {}
  t1 += b.report("int => int:") do
    n.times {|x| h[x] = x }
  end
  t1 += b.report("int => [int]:") do
    n.times {|x| h[x] = [x] }
  end
  t1 += b.report(".to_sym => .to_s:") do
    n.times {|x| h[x.to_s.to_sym] = x.to_s }
  end
 
  puts ":sym => str 1.000.000:"
  t2 = b.report("Each:") do
    h.each {|k,v| k; v}
  end
 
  t2 += b.report("each_key:") do
    h.each_key {|k| k}
  end
 
  t2 += b.report("each_value:") do
    h.each_value {|v| v}
  end
 
  t2 += b.report("read by key:") do
    h.each {|k,v| h[k]}
  end
 
  h1=h
 
  t3 = b.report("5 invert:") do
    5.times { h.invert }
  end
 
  h=h1
  t3 += b.report("1000 shift:") do
    1000.times { h.shift }
  end
 
  h=h1
  t3 += b.report("1000 delete:") do
    1000.times {|x| h.delete(x) }
  end
 
  h=h1
  t3 += b.report("all delete:") do
    h.each_key {|k| h.delete(k)}
  end
 
  [t1+t2+t3]
end

Ruby 1.8.6

C:>ruby bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            0.851000   0.030000   0.881000 (  0.881320)
int => int:            1.021000   0.090000   1.111000 (  1.141710)
int => [int]:          0.791000   0.030000   0.821000 (  0.841260)
.to_sym => .to_s:      6.400000   0.080000   6.480000 (  6.509750)
:sym => str 1.000.000:
Each:                  2.553000   0.090000   2.643000 (  2.684020)
each_key:              0.481000   0.000000   0.481000 (  0.490735)
each_value:            0.491000   0.000000   0.491000 (  0.490735)
read by key:           2.854000   0.040000   2.894000 (  2.904350)
5 invert:             31.185000   0.791000  31.976000 ( 32.438585)
1000 shift:           37.604000   0.010000  37.614000 ( 38.177180)
1000 delete:           0.000000   0.000000   0.000000 (  0.000000)
all delete:            1.442000   0.010000   1.452000 (  1.452175)
Total:                85.673000   1.171000  86.844000 ( 88.011820)

Ruby 1.9.1

C:>ruby bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            0.721000   0.100000   0.821000 (  0.841260)
int => int:            0.590000   0.140000   0.730000 (  0.731095)
int => [int]:          0.391000   0.020000   0.411000 (  0.440660)
.to_sym => .to_s:      4.957000   0.320000   5.277000 (  5.327980)
:sym => str 1.000.000:
Each:                  0.691000   0.050000   0.741000 (  0.801200)
each_key:              0.461000   0.000000   0.461000 (  0.470705)
each_value:            0.500000   0.000000   0.500000 (  0.520780)
read by key:           1.022000   0.011000   1.033000 (  1.051575)
5 invert:             33.939000   1.071000  35.010000 ( 35.603325)
1000 shift:           48.800000   0.020000  48.820000 ( 49.474100)
1000 delete:           0.000000   0.000000   0.000000 (  0.000000)
all delete:            1.432000   0.010000   1.442000 (  1.472205)
Total:                93.504000   1.742000  95.246000 ( 96.734885)

JRuby

C:>jruby bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            2.113000   0.000000   2.113000 (  1.933000)
int => int:            0.721000   0.000000   0.721000 (  0.721000)
int => [int]:          1.022000   0.000000   1.022000 (  1.022000)
.to_sym => .to_s:     34.641000   0.000000  34.641000 ( 34.641000)
:sym => str 1.000.000:
Each:                  0.541000   0.000000   0.541000 (  0.541000)
each_key:              0.191000   0.000000   0.191000 (  0.191000)
each_value:            0.240000   0.000000   0.240000 (  0.240000)
read by key:           0.761000   0.000000   0.761000 (  0.761000)
5 invert:            Error: Your application used more memory than the safety cap.

IronRuby

C:>ir bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            2.183139   0.030043   2.213182 (  2.213315)
int => int:            1.211742   0.030043   1.241786 (  1.291935)
int => [int]:          1.712462   0.010014   1.722477 (  1.732595)
.to_sym => .to_s:      7.460728   0.120173   7.580901 (  7.781655)
:sym => str 1.000.000:
Each:                  1.111598   0.000000   1.111598 (  1.131695)
each_key:              0.340490   0.000000   0.340490 (  0.350525)
each_value:            0.330475   0.010014   0.340490 (  0.350525)
read by key:           1.021469   0.000000   1.021469 (  1.031545)
5 invert:              5.588035   0.350504   5.938539 (  7.381055)
1000 shift:            0.020029   0.000000   0.020029 (  0.190285)
1000 delete:           0.020029   0.000000   0.020029 (  0.080120)
all delete:            1.031483   0.020029   1.051512 (  1.061590)
Total:                22.031680   0.570821  22.602501 ( 24.596839)

Summary:

Ruby 1.8.6
Total: 85.673000 1.171000 86.844000 ( 88.011820)
Ruby 1.9.1
Total: 93.504000 1.742000 95.246000 ( 96.734885)
Jruby
*** Out of memory ***
Ironruby
Total: 22.031680 0.570821 22.602501 ( 24.596839)

Comments: JRuby has not concluded the test to an error of "out of memory", the java process has come to occupy 600Mb in the test where you make an {}.invert, reaching the limit of the JVM. The least greedy of memory was ruby 1.8.6 with "only" 300Mb while 1.9.1 was almost fixed at 530Mb, not increased, as if there was a wall. There was not, instead, for Ironruby that has exceeded the barrier of 600Mb but with great results, the inversion executed in 5 seconds, compared with 33 of 1.9.1!

The same test with 2Gb of ram:

Ruby 1.8.6

C:>ruby bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            0.811000   0.060000   0.871000 (  0.871252)
int => int:            1.081000   0.050000   1.131000 (  1.141642)
int => [int]:          0.751000   0.030000   0.781000 (  0.801152)
.to_sym => .to_s:      6.299000   0.130000   6.429000 (  6.469302)
:sym => str 1.000.000:
Each:                  2.474000   0.060000   2.534000 (  2.553672)
each_key:              0.491000   0.000000   0.491000 (  0.490706)
each_value:            0.500000   0.000000   0.500000 (  0.500720)
read by key:           2.834000   0.080000   2.914000 (  2.964263)
5 invert:             30.454000   0.992000  31.446000 ( 31.885849)
1000 shift:           38.285000   0.020000  38.305000 ( 38.765743)
1000 delete:           0.000000   0.000000   0.000000 (  0.000000)
all delete:            1.472000   0.010000   1.482000 (  1.512174)
Total:                85.452000   1.432000  86.884000 ( 87.956475)

Ruby 1.9.1

C:>ruby bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            0.701000   0.090000   0.791000 (  0.790640)
int => int:            0.661000   0.100000   0.761000 (  0.780631)
int => [int]:          0.360000   0.020000   0.380000 (  0.390316)
.to_sym => .to_s:      4.947000   0.380000   5.327000 (  5.524472)
:sym => str 1.000.000:
Each:                  0.631000   0.041000   0.672000 (  0.690558)
each_key:              0.451000   0.000000   0.451000 (  0.450365)
each_value:            0.490000   0.000000   0.490000 (  0.490397)
read by key:           0.892000   0.030000   0.922000 (  0.940761)
5 invert:             27.760000   6.299000  34.059000 ( 34.547961)
1000 shift:           48.289000   0.020000  48.309000 ( 48.789488)
1000 delete:           0.000000   0.000000   0.000000 (  0.000000)
all delete:            1.072000   0.370000   1.442000 (  1.441166)
Total:                86.254000   7.350000  93.604000 ( 94.836755)

JRuby

C:>jruby bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            2.261000   0.000000   2.261000 (  2.081000)
int => int:            0.760000   0.000000   0.760000 (  0.760000)
int => [int]:          1.051000   0.000000   1.051000 (  1.051000)
.to_sym => .to_s:     35.694000   0.000000  35.694000 ( 35.694000)
:sym => str 1.000.000:
Each:                  0.481000   0.000000   0.481000 (  0.481000)
each_key:              0.190000   0.000000   0.190000 (  0.190000)
each_value:            0.190000   0.000000   0.190000 (  0.190000)
read by key:           0.701000   0.000000   0.701000 (  0.701000)
5 invert:            Error: Your application used more memory than the safety cap.

C:>jruby bench_hsh.rb -J-Xmx1500m
                           user     system      total        real
Write 1.000.000:
int => str:            1.612000   0.000000   1.612000 (  1.562000)
int => int:            0.842000   0.000000   0.842000 (  0.842000)
int => [int]:          1.101000   0.000000   1.101000 (  1.101000)
.to_sym => .to_s:     35.341000   0.000000  35.341000 ( 35.341000)
:sym => str 1.000.000:
Each:                  0.431000   0.000000   0.431000 (  0.431000)
each_key:              0.180000   0.000000   0.180000 (  0.180000)
each_value:            0.190000   0.000000   0.190000 (  0.200000)
read by key:           0.701000   0.000000   0.701000 (  0.691000)
5 invert:             11.076000   0.000000  11.076000 ( 11.076000)
1000 shift:            0.000000   0.000000   0.000000 (  0.000000)
1000 delete:           0.000000   0.000000   0.000000 (  0.000000)
all delete:            0.912000   0.000000   0.912000 (  0.912000)
Total:                52.385999   0.000000  52.385999 ( 52.335999)

IronRuby

C:>ir bench_hsh.rb
                           user     system      total        real
Write 1.000.000:
int => str:            2.283283   0.090130   2.373413 (  2.363398)
int => int:            1.201728   0.030043   1.231771 (  1.241786)
int => [int]:          1.792578   0.020029   1.812606 (  1.862678)
.to_sym => .to_s:      7.340555   0.120173   7.460728 (  7.590915)
:sym => str 1.000.000:
Each:                  1.091570   0.000000   1.091570 (  1.141642)
each_key:              0.310446   0.000000   0.310446 (  0.310447)
each_value:            0.360518   0.000000   0.360518 (  0.360518)
read by key:           1.001440   0.000000   1.001440 (  1.001440)
5 invert:              4.816926   0.070101   4.887027 (  4.917070)
1000 shift:            0.010014   0.000000   0.010014 (  0.010014)
1000 delete:           0.010014   0.000000   0.010014 (  0.010015)
all delete:            1.061526   0.000000   1.061526 (  1.071541)
Total:                21.280600   0.330475  21.611075 ( 21.881464)

Summary:

Ruby 1.8.6
Total: 85.452000 1.432000 86.884000 ( 87.956475)
Ruby 1.9.1
Total: 86.254000 7.350000 93.604000 ( 94.836755)
Jruby
Total: 52.385999 0.000000 52.385999 ( 52.335999)
Ironruby
Total: 21.280600 0.330475 21.611075 ( 21.881464)

Comments: JRuby hangs again and always with a process of 600Mb, so I increased the cap to 1.5 Gb memory allowing it to finish even with a good result but the process has reached monstrous size, has passed 1Gb. Ruby 1.8.6 again the least greedy with 300Mb and again 530Mb for the 1.9.1. Both have high time with the invert method and in particular with the shift, contrary for the two others. Ironruby still the fastest with a process that detects an huge amount of memory, coming to allocate almost 700Mb.
Deleting the elements has always been fast even if they increased the number, maybe I did not check well that were not performed in a single call, however, I shall from this result.

Conclusions:

Each interpreter has its areas where it excels, Ironruby when he does, he does well, outdistanced the others. JRuby also has taken its triumphs, maybe I had a little more expectations from version 1.9.1 although I believe it is not a demerit but were the others to improve much in recent months. I was very impressed by the version 1.8.6 with this "new" patch 368 with, or better, a different compiled (mingw32), that much improved performance by shortening the distances with the new 1.9, while not having a VM. It is also the most balanced without high or low and to consider that there is still the most used, and more compatible with the largest number of buds.

I hope this article was interesting, the next will include a comparison with also python 2.6.2 and the newest 3.1.1 which is very very fast. Hola!

 
Comments Off

Posted in IronRuby .NET, JRuby, Ruby