Rich Freeman on 5 Jul 2013 04:54:27 -0700 |
[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]
Re: [PLUG] multiprocessing on Python |
On Thu, Jul 4, 2013 at 3:37 PM, Christopher Ochsenreither <c.g.ochsenreither@gmail.com> wrote: > Someone at the PLUG central meeting was asking about distributing a Python > process over multiple cores. I thought this article might be of interest: I was messing around with those modules when I was trying to move my git repository analyzer map/reduce solution out of hadoop. I struggled to get them to work properly and saturate my CPUs, and ended up just using GNU parallel instead. For cases where you are just processing a fixed input that can be split and processed in any order (like map/reduce) GNU parallel works very well. Since i was running multiple map/reduce iterations it also allowed me to run map and reduce states in parallel by pipe (the sort step obviously limited the ability to completely pipe all the stages effectively). However, if your parallelism needs are limited to a subset of your program's functionality those python modules probably are the best place to start... Rich ___________________________________________________________________________ Philadelphia Linux Users Group -- http://www.phillylinux.org Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce General Discussion -- http://lists.phillylinux.org/mailman/listinfo/plug