[IPython-User] distributing jobs

[MinRK]

On Thu, Jun 28, 2012 at 1:45 PM, Robert Nishihara <robertnishihara@gmail.com
> wrote:

> Say I start 100 engines with
>
>     ipcontroller --profile=sge
>     ipcluster engines -n 100 --profile=sge
>
> And I want to map a function over a list of length 100 as below
>
>     >> from IPython import parallel
>     >> rc = parallel.Client(profile='sge')
>     >> dview = rc[:]
>     >> dview.map_sync(f, range(100))
>
> Since there are 100 engines and 100 function evaluations to do, is it the
> case that map_sync will send one evaluation to each engine?
>

Yes.  DirectView.map makes at most one apply call on each engine with the
appropriate partition of the list.


>
> -Robert
>
> _______________________________________________
> IPython-User mailing list
> IPython-User@scipy.org
> http://mail.scipy.org/mailman/listinfo/ipython-user
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/ipython-user/attachments/20120628/67814fd5/attachment.html