From: Daniel Berger <djberg96@...>
Date: 2008-07-11T11:39:44+09:00
Subject: Re: threadify-0.0.1



On Jul 1, 2:04 pm, ara howard <ara.t.how...@gmail.com> wrote:
> this one's for you charlie ;-)
>
> NAME
>    threadify.rb
>
> SYNOPSIS
>    enumerable = %w( a b c d )
>    enumerable.threadify(2){ 'process this block using two worker  
> threads' }
>
> DESCRIPTION
>    threadify.rb makes it stupid easy to process a bunch of data using  
> 'n'
>    worker threads
>
> INSTALL
>    gem installthreadify
>
> URI
>    http://rubyforge.org/projects/codeforpeople
>
> SAMPLES
>
>    <========< sample/a.rb >========>
>
>    ~ > cat sample/a.rb
>
>      require 'open-uri'
>      require 'yaml'
>
>      require 'rubygems'
>      require 'threadify'
>
>      uris =
>        %w(
>          http://google.com
>          http://yahoo.com
>          http://rubyforge.org
>          http://ruby-lang.org
>          http://kcrw.org
>          http://drawohara.com
>          http://codeforpeople.com
>        )
>
>      time 'withoutthreadify' do
>        uris.each do |uri|
>          body = open(uri){|pipe| pipe.read}
>        end
>      end
>
>      time 'withthreadify' do
>        uris.threadifydo |uri|
>          body = open(uri){|pipe| pipe.read}
>        end
>      end
>
>      BEGIN {
>        def time label
>          a = Time.now.to_f
>          yield
>        ensure
>          b = Time.now.to_f
>          y label => (b - a)
>        end
>      }
>
>    ~ > ruby sample/a.rb
>
>      ---
>      withoutthreadify: 7.41900205612183
>      ---
>      withthreadify: 3.69886112213135

Pretty cool. I tried it with file-find. Here was the code:

require 'file/find'
require 'threadify'

rule = File::Find.new(
   :pattern => "*.rb",
   :path    => "C:\\ruby"
)

start = Time.now

rule.find.threadify(10){ |f|
   p f
}

p start
p Time.now

Without threadify, it took 1:40 on my laptop. With threadify(10) it
dropped to 44 seconds.

I think I'll add a "threads" option directly, and borrow some of your
code. :)

Thanks,

Dan