# \[ANN\] forkoff-1.1.0 **URL:** \ **Category:** ruby-talk\ **Created:** [12 October 2009 03:58 UTC](https://rubytalk.org/t/ann-forkoff-1-1-0/55765 "2009-10-12T03:58:26Z")\ **Posts on this page:** 1\ **Page:** 1
**Author:** ![a11](https://yyz1.discourse-cdn.com/flex029/user_avatar/rubytalk.org/a11/32/8169_2.png) [@a11](https://rubytalk.org/u/a11)\ **Post date:** [12 October 2009 03:58 UTC](https://rubytalk.org/t/ann-forkoff-1-1-0/55765/1 "2009-10-12T03:58:26Z")
NAME   forkoff SYNOPSIS   brain-dead simple parallel processing for ruby URI   [http://rubyforge.org/projects/codeforpeople](http://rubyforge.org/projects/codeforpeople)   [http://github.com/ahoward/forkoff](http://github.com/ahoward/forkoff) INSTALL   gem install forkoff DESCRIPTION   forkoff works for any enumerable object, iterating a code block to run in a   child process and collecting the results. forkoff can limit the number of   child processes which is, by default, 2. SAMPLES   \<========\< samples/a.rb \>========\>   ~ \> cat samples/a.rb     # forkoff makes it trivial to do parallel processing with ruby, the following     # prints out each word in a separate process > **ยทยทยท** > > # > >       require 'forkoff' > >       %w( hey you ).forkoff!{|word| puts "#{ word } from # > { Process.pid }"} > >   ~ \> ruby samples/a.rb > >     hey from 7907 >     you from 7908 > >   \<========\< samples/b.rb \>========\> > >   ~ \> cat samples/b.rb > >     # for example, this takes only 4 seconds or so to complete (8 > iterations >     # running in two processes = twice as fast) >     # > >       require 'forkoff' > >       a = Time.now.to\_f > >       results = >         (0..7).forkoff do |i| >           sleep 1 >           i \*\* 2 >         end > >       b = Time.now.to\_f > >       elapsed = b - a > >       puts "elapsed: #{ elapsed }" >       puts "results: #{ results.inspect }" > >   ~ \> ruby samples/b.rb > >     elapsed: 4.19184589385986 >     results: [0, 1, 4, 9, 16, 25, 36, 49] > >   \<========\< samples/c.rb \>========\> > >   ~ \> cat samples/c.rb > >     # forkoff does \*NOT\* spawn processes in batches, waiting for each > batch to >     # complete. rather, it keeps a certain number of processes busy > until all >     # results have been gathered. in otherwords the following will > ensure that 3 >     # processes are running at all times, until the list is complete. > note that >     # the following will take about 3 seconds to run (3 sets of 3 @ 1 > second). >     # > >     require 'forkoff' > >     pid = Process.pid > >     a = Time.now.to\_f > >     pstrees = >       %w( a b c d e f g h i ).forkoff! :processes =\> 3 do |letter| >         sleep 1 >         { letter =\> ` pstree -l 2 #{ pid } ` } >       end > >     b = Time.now.to\_f > >     puts >     puts "pid: #{ pid }" >     puts "elapsed: #{ b - a }" >     puts > >     require 'yaml' > >     pstrees.each do |pstree| >       y pstree >     end > >   ~ \> ruby samples/c.rb > >     pid: 7922 >     elapsed: 3.37899208068848 > >     --- >     a: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>-+- 07923 ahoward ruby -Ilib samples/c.rb >        \>-+- 07924 ahoward (ruby) >        \-+- 07925 ahoward ruby -Ilib samples/c.rb > >     --- >     b: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>-+- 07923 ahoward ruby -Ilib samples/c.rb >        \>-+- 07924 ahoward ruby -Ilib samples/c.rb >        \-+- 07925 ahoward ruby -Ilib samples/c.rb > >     --- >     c: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>-+- 07923 ahoward ruby -Ilib samples/c.rb >        \>-+- 07924 ahoward (ruby) >        \-+- 07925 ahoward ruby -Ilib samples/c.rb > >     --- >     d: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>-+- 07932 ahoward ruby -Ilib samples/c.rb >        \>--- 07933 ahoward ruby -Ilib samples/c.rb >        \--- 07934 ahoward ruby -Ilib samples/c.rb > >     --- >     e: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>--- 07932 ahoward (ruby) >        \>-+- 07933 ahoward ruby -Ilib samples/c.rb >        \-+- 07934 ahoward (ruby) > >     --- >     f: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>--- 07932 ahoward (ruby) >        \>-+- 07933 ahoward ruby -Ilib samples/c.rb >        \-+- 07934 ahoward ruby -Ilib samples/c.rb > >     --- >     g: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>-+- 07941 ahoward ruby -Ilib samples/c.rb >        \>--- 07942 ahoward ruby -Ilib samples/c.rb >        \--- 07943 ahoward ruby -Ilib samples/c.rb > >     --- >     h: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>-+- 07941 ahoward (ruby) >        \>-+- 07942 ahoward ruby -Ilib samples/c.rb >        \--- 07943 ahoward ruby -Ilib samples/c.rb > >     --- >     i: | >       -+- 07922 ahoward ruby -Ilib samples/c.rb >        \>--- 07942 ahoward (ruby) >        \-+- 07943 ahoward ruby -Ilib samples/c.rb > >   \<========\< samples/d.rb \>========\> > >   ~ \> cat samples/d.rb > >     # forkoff supports two strategies of reading the result from the > child: via >     # pipe (the default) or via file. you can select which to use > using the >     # :strategy option. >     # > >       require 'forkoff' > >       %w( hey you guys ).forkoff :strategy =\> :file do |word| >         puts "#{ word } from #{ Process.pid }" >       end > >   ~ \> ruby samples/d.rb > >     hey from 7953 >     you from 7954 >     guys from 7955 > > HISTORY >   1.1.0 >     - move to a model with one work queue and signals sent from > consumers to >     producer to noitify ready state. this let's smaller jobs race > through a >     single process even while a larger job may have one sub-process > bound up. >     incorporates a fix from [http://github.com/fredrikj/forkoff](http://github.com/fredrikj/forkoff) which > meant >     some processes would lag behind when jobs didn't have similar > execution >     times. > >   1.0.0 >     - move to github > >   0.0.4 >     - code re-org >     - add :strategy option >     - default number of processes is 2, not 8 > >   0.0.1 > >     - updated to use producer threds pushing onto a SizedQueue for > each consumer >       channel. in this way the producers do not build up a massize > parllel data >       structure but provide data to the consumers only as fast as they > can fork >       and proccess it. basically for a 4 process run you'll end up > with 4 >       channels of size 1 between 4 produces and 4 consumers, each > consumer is a >       thread popping of jobs, forking, and yielding results. > >     - removed use of Queue for capturing the output. now it's simply > an array >       of arrays which removed some sync overhead. > >     - you can configure the number of processes globally with > >         Forkoff.default['proccess'] = 4 > >     - you can now pass either an options hash > >         forkoff( :processes =\> 2 ) ... > >       or plain vanilla number > >         forkoff( 2 ) ... > >       to the forkoff call > >     - default number of processes is 8, not 2 > >   0.0.0 > >     initial version > > enjoy.