[mpiwg-p2p] Schedule for Dec 2014 face-to-face meeting in San Jose

Jim Dinan james.dinan at gmail.com
Thu Dec 4 09:16:33 CST 2014

Hi Dan,

I agree that it would be helpful to review the wildcards info key to gather
more inputs and make sure that folks in the P2P WG are comfortable with the
proposal.  The process could have been better, most of that is my fault,
but let's not miss this opportunity to be the first MPI 4.0 ticket.  :)  If
possible, a little WG time to review other info keys and prioritize would
be helpful so that we can pick up and move forward the other great ideas
that are captured on the ticket.  We definitely do have other info keys
that belong in this working group.

Safe travels and see you next week!

On Thu, Dec 4, 2014 at 9:34 AM, Daniel Holmes <dholmes at epcc.ed.ac.uk> wrote:

> Hi all,
> I propose the following as the initial schedule for the P2P WG time at the
> face-to-face WG meeting in San Jose:
> 09:00 Receive w/reduce (30 minutes)
> 09:30 Allocate receive (30 minutes)
> 10:00 -----Break------ (30 minutes)
> 10:30 Wildcards ticket (15 minutes)
> 10:45 Streams/channels (75 minutes)
> 12:00 -----Lunch------ (60 minutes)
> Jeff Hammond will be presenting some new information about "receive with
> reduce" and "allocate receive".
> The "wildcards ticket" is intended to be a short time to make a decision
> on whether or not the P2P WG should pick up and drive the spin-off INFO
> keys ticket (#461) that proposes a couple of "no wildcards" keys and
> whether or not to offer to take on other proposed keys (as long as they are
> P2P-related) as lower priority items (e.g. "only_any_source" and
> "only_any_tag").
> Finally, we still have a lot of material regarding streams/channels to
> discuss. (Note-to-self: Takeshi Nanri, Kyushu Univ., Japan on channels in
> ML and Xiaoyi Lu, Ohio State Univ., USA on DataMPI.)
> I realise that many P2P WG members are also Hybrid WG members and the
> parallel track in the afternoon of Tuesday 9th is not ideal for us. I am
> proposing that we restrict our WG time to just the two morning sessions to
> avoid the timetabling clash. We can make a decision during lunch, based on
> our progress through the agenda by that time, about whether the afternoon
> P2P session is used or not.
> Cheers,
> Dan.
> --
> 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-p2p mailing list
> mpiwg-p2p at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-p2p
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpi-forum.org/pipermail/mpiwg-p2p/attachments/20141204/e28c75d3/attachment-0001.html>

More information about the mpiwg-p2p mailing list