MathGroup Archive 2011

[Date Index] [Thread Index] [Author Index]

Search the Archive

Nested ParallelTable

  • To: mathgroup at smc.vnet.net
  • Subject: [mg120646] Nested ParallelTable
  • From: Ivan <ivan.kassal at gmail.com>
  • Date: Mon, 1 Aug 2011 06:58:17 -0400 (EDT)
  • Delivered-to: l-mathgroup@mail-archive0.wolfram.com

Hi all,

If I run, on an eight-core machine and with Mathematica 7,
ParallelTable[Pause[1];$KernelID, {i,1,2},{j,1,2}]
I get
{{1,1},{2,2}}
And the evaluation takes 2 seconds. What I would expect to happen is
to get
{{1,2},{3,4}}
and the evaluation taking only 1 second. That would seem like the
natural results, considering that
ParallelTable[Pause[1];$KernelID, {i,1,4}]
indeed returns
{1,2,3,4}

Is there any way to get the parallelization to work on all levels of
the table, not just the top level?

Thanks!




  • Next by Date: Re: Local scoping of pattern names in rules--or not
  • Next by thread: Re: Local scoping of pattern names in rules--or not