[mpiwg-hybridpm] Fwd: [mpiwg-persistence] HACC WG This Wednesday

Joseph Schuchart schuchart at icl.utk.edu
Tue Sep 28 18:00:02 CDT 2021


Sorry, the email went to the persistence mailing list only...


-------- Forwarded Message --------
Subject: 	Re: [mpiwg-persistence] HACC WG This Wednesday
Date: 	Tue, 28 Sep 2021 18:59:05 -0400
From: 	Joseph Schuchart <schuchart at icl.utk.edu>
To: 	mpiwg-persistence at lists.mpi-forum.org



Hi all,

On 9/27/21 11:40 AM, Jim Dinan via mpiwg-persistence wrote:
> Hi All,
>
> This Wednesday, the HACC WG will meet and discuss:
>
> 1) Message ordering in the case of multiple threads (#117).

I put my proposal for turning the disputed paragraph into a short AtoU 
up at [1] so we can work on it during the meeting. If someone has a 
better idea for how to organize this out of the eye of the broader forum 
please let me know.

> 2) The draft continuations proposal.

I'm attaching the current PDF (which is similar to the one I sent out a 
while ago). The github diff can be found at [2]. We had some internal 
discussions about changing the current proposal and found that the lack 
of a compelling use-case for multiple continuations doesn't help in 
justifying the added complexity and overhead. We will go with the 
current version of it.

See you tomorrow,
Joseph

[1] https://github.com/mpi-forum/mpi-standard/pull/627
[2] 
https://github.com/mpi-forum/mpi-standard/compare/mpi-4.x...devreal:mpi-continuations

>
> I would appreciate if folks discussing on #117 could bring proposals 
> for how to clarify this text. Joseph, could you also please share your 
> latest draft for the continuations proposal?
>
> Thanks,
>  ~Jim.
>
> _______________________________________________
> mpiwg-persistence mailing list
> mpiwg-persistence at lists.mpi-forum.org
> https://lists.mpi-forum.org/mailman/listinfo/mpiwg-persistence


-------------- next part --------------
A non-text attachment was scrubbed...
Name: mpi40-report-continuations.pdf
Type: application/pdf
Size: 212993 bytes
Desc: not available
URL: <http://lists.mpi-forum.org/pipermail/mpiwg-hybridpm/attachments/20210928/2750a664/attachment-0001.pdf>


More information about the mpiwg-hybridpm mailing list