Subject: Re: [boost] [mpi] Automatically Cascading/Nested Reduce?
From: John Phillips (phillips_at_[hidden])
Date: 2009-04-21 09:16:12
Dean Michael Berris wrote:
> Hi John,
> On Tue, Apr 21, 2009 at 7:46 AM, John Phillips
> <phillips_at_[hidden]> wrote:
>> What MPI implementation are you using?
> I'm using OpenMPI version 1.2.6 on Gentoo Linux.
> I haven't looked at whether OpenMPI does what I'm looking for that's
> why I was asking if it might be worth implementing at the Boost.MPI
> layer. Although it would be nice if OpenMPI did it for me in the first
> place. ;-)
Their current version is 1.3.1, and I know the transition to 1.3
included a number of performance optimizations in the code. They also
have a fairly active, and quite responsive user list, with good
involvement from the developers. So, as a first step, I'd suggest
checking with them. I don't think Doug's intent when he submitted
Boost.MPI was to have it evolve into a MPI implementation, so I don't
know that he would want to add something like this. (He is, obviously,
the one to ask to be sure.)
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk