[mpiwg-rma] Fwd: [MPI Forum] #456: MPI 3.0 Errata: MPI RMA synchronization for MPI shared memory

[Rolf Rabenseifner]

Dear members of the RMA WG,

after first partial reviews by Jeff Hammond, Jed Brown and Dave Goodell,
I could fill the TO-DO areas and hopefully finalize #456. 
https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/456

Just in time before the Sep.1 deadline.
It is based on the result of the last telecon and the 
general rules shared memory barriers.

The way of describing is based on the goal to define the 
outcome and not the implementation of the MPI library.
The solution touches only shared memory windows.

I hope that the RMA WG can discuss it before or in Kobe.

In my opinion, MPI-3.1 should not be finalized before 
this problem is resolved.
It is an open topic on the RMA agenda since 2 years,
since the request for clarification from 11-Sep-2012 
about "MPI_Win_allocate_shared and synchronization functions".

Best regards
Rolf

----- Original Message -----
> From: "Rolf Rabenseifner" <rabenseifner at hlrs.de>
> To: "MPI WG Remote Memory Access working group" <mpiwg-rma at lists.mpi-forum.org>
> Sent: Thursday, August 14, 2014 1:00:55 AM
> Subject: [mpiwg-rma] Fwd: [MPI Forum] #456: MPI 3.0 Errata: MPI RMA synchronization for MPI shared memory
> 
> Dear Bill and whole RMA working Group,
> 
> I tried to bring all my knowledge together to solve the
> "shared memory synchronization problem".
> 
> I know, that you Bill also wanted to solve this problem.
> Please take my ticket
> https://svn.mpi-forum.org/trac/mpi-forum-web/ticket/456
> as a basis or stone quarry or ... .
> 
> For me, it is important to really find all locations that must
> be touched to solve the Problem and to have a consistent solution.
> 
> I hope that I found all of them and my solution is consistent.
> 
> The solution is free of technical implementation internals,
> i.e., the local memory barriers are only mentioned in a rationale.
> 
> I hope you all enjoy the proposal and it is a basis for
> further improvements.
> 
> Best bregards
> Rolf
> 
> 
> --
> Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner at hlrs.de
> High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530
> University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832
> Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner
> Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307)
> _______________________________________________
> mpiwg-rma mailing list
> mpiwg-rma at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma
> 

-- 
Dr. Rolf Rabenseifner . . . . . . . . . .. email rabenseifner at hlrs.de
High Performance Computing Center (HLRS) . phone ++49(0)711/685-65530
University of Stuttgart . . . . . . . . .. fax ++49(0)711 / 685-65832
Head of Dpmt Parallel Computing . . . www.hlrs.de/people/rabenseifner
Nobelstr. 19, D-70550 Stuttgart, Germany . . . . (Office: Room 1.307)