[mpiwg-hybridpm] Info Keys Ticket

[Jim Dinan]

Hi Dan,

Sure, let's gather what we can.  For these info keys, the benefit is fairly
straightforward.  For communicators with wildcard restrictions you can use
a structure that is much more efficient to search than a single ordered
list.  In addition, when you have multiple communication channels (e.g.
shared memory, multiple rails, etc) no_any_source removes a lot of
complications and overheads.

 ~Jim.

On Thu, Nov 20, 2014 at 2:09 PM, Daniel Holmes <dholmes at epcc.ed.ac.uk>
wrote:

>  Hi All,
>
> One of the objections last time was that these keys seem like a good idea
> for performance but there's no implementation/proof. We'll need to collate
> some of the existing research that shows this gives a performance benefit
> in order to address that objection.
>
> Cheers,
> Dan.
>
>
> On 20/11/2014 13:03, Jim Dinan wrote:
>
> Hi All,
>
>  I would like to bring forward the following ticket on communicator info
> keys for a formal reading at the upcoming meeting:
> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/461.
>
>  This ticket contains the no_any_tag and no_any_source info keys from
> #381 on which we were able to reach consensus.  There are lots of other
> good ideas on #318, and we will continue to work on moving those ahead.
>
>  Please send feedback.  The deadline for changes to the proposal is
> Sunday.
>
>  Thanks,
>  ~Jim.
>
>
> _______________________________________________
> mpiwg-hybridpm mailing listmpiwg-hybridpm at lists.mpi-forum.orghttp://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-hybridpm
>
>
> --
> Dan Holmes
> Applications Consultant in HPC Research
> EPCC, The University of Edinburgh
> James Clerk Maxwell Building
> The Kings Buildings
> Peter Guthrie Tait Road
> Edinburgh
> EH9 3FD
> T: +44(0)131 651 3465
> E: dholmes at epcc.ed.ac.uk
>
> *Please consider the environment before printing this email.*
>
>
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
> _______________________________________________
> mpiwg-hybridpm mailing list
> mpiwg-hybridpm at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-hybridpm
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpi-forum.org/pipermail/mpiwg-hybridpm/attachments/20141120/f48908e8/attachment-0001.html>