[mpiwg-hybridpm] Meeting Tomorrow

[Jim Dinan]

To register my $0.02 -- A model that doesn't allow the application to order
MPI operations via application-level synchronization is substantially less
useful than one that does. If we're worried about performance impacts of
this (none exist today), then we can either add a new threading model or
use info assertions to allow applications to opt-in to threaded
communication being un-orderable.

On Tue, May 9, 2023 at 10:41 AM Holmes, Daniel John <
daniel.john.holmes at intel.com> wrote:

> Hi all,
>
>
>
> Unfortunately, I am double-booked for tomorrow’s HACC WG time slot – so my
> answer to the implied question below is “not yet”.
>
>
>
> The “logically concurrent isn’t” issue #117 is now accepted and merged
> into MPI-4.1 (take a moment to celebrate!) – but it just says “here be
> dragons”.
>
>
>
> Do we care enough to defeat those dragons?
>
>
>
> Argument FOR: as systems become more heterogeneous, an MPI process is
> likely to abstractly “contain” more semi-independent HW units that will
> want to communicate with other MPI processes, which will result in lots of
> logically concurrent MPI communication operations – exactly the territory
> in which these dragons live.
>
>
>
> Argument AGAINST: we’ve been throwing brave warriors into this particular
> dragon fire for about a decade and we’ve only now convinced ourselves that
> the dragons do, in fact, exist. How many more volunteers do we have and do
> they have sufficiently pointy sticks?
>
>
>
> Best wishes,
>
> Dan.
>
>
>
> *From:* mpiwg-hybridpm <mpiwg-hybridpm-bounces at lists.mpi-forum.org> *On
> Behalf Of *Jim Dinan via mpiwg-hybridpm
> *Sent:* 09 May 2023 15:15
> *To:* Hybrid working group mailing list <
> mpiwg-hybridpm at lists.mpi-forum.org>
> *Cc:* Jim Dinan <james.dinan at gmail.com>
> *Subject:* [mpiwg-hybridpm] Meeting Tomorrow
>
>
>
> Hi All,
>
>
>
> We had to reschedule the topic planned for tomorrow's meeting, so the
> agenda is now open. Please let me know if you have a topic you'd like to
> discuss. If we don't have a topic ahead of time, we will cancel.
>
>
>
> Thanks,
>
>  ~Jim.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpi-forum.org/pipermail/mpiwg-hybridpm/attachments/20230510/24227438/attachment.html>