[Mpi3-rma] [Mpi-comments] Freeing memory allocated in MPI_WIN_ALLOCATE_SHARED

[William Gropp]

Thanks.  I like this proposal.  Should I include it in the draft update?  Let me know ASAP, as we are running out of time to make changes.

Bill

William Gropp
Director, Parallel Computing Institute
Deputy Director for Research
Institute for Advanced Computing Applications and Technologies
Paul and Cynthia Saylor Professor of Computer Science
University of Illinois Urbana-Champaign



On Sep 11, 2012, at 11:29 AM, Hubert Ritzdorf wrote:

> Hi Torsten,
> 
> I agree. What's about this proposal for which Bill was asking:
> 
> Page 407, after Line 46 (relative to mpi3,0_draft2.pdf):
> \mpifunc{MPI\_WIN\_FREE} will free the window memory that was allocated in
> \mpifunc{MPI\_WIN\_ALLOCATE}.  
> 
> Page 409, after Line 29:
> \mpifunc{MPI\_WIN\_FREE} will free the window memory that was allocated in
> \mpifunc{MPI\_WIN\_ALLOCATE\_SHARED}. 
> 
> Page 415,  Line 5:
> Replace 
> MPI_WIN_FREE will free the window memory that was allocated in MPI_WIN_ALLOCATE.
> by
> \mpifunc{MPI\_WIN\_FREE} might detach or free the window memory. The specific window creation and
> attach functions specify effects on the window memory by  \mpifunc{MPI\_WIN\_FREE}.
> 
> Page 415, Line 6-8:
> To be consistent, the text "Freeing a window that was created with ..." should be removed and
> mainly moved to:
> 
> Page 414, after Line 3:
> 
> \mpifunc{MPI\_WIN\_FREE)  will detach all associated memory; i.e. it has the same effect
> as if all attached memory was detached by calls to \mpifunc{MPI\_WIN\_DETACH}.
> 
> Hubert
> 
> ________________________________________
> From: mpi3-rma-bounces at lists.mpi-forum.org [mpi3-rma-bounces at lists.mpi-forum.org] on behalf of Torsten Hoefler [htor at illinois.edu]
> Sent: Tuesday, September 11, 2012 9:14 AM
> To: MPI 3.0 Remote Memory Access working group
> Subject: Re: [Mpi3-rma] [Mpi-comments] Freeing memory allocated in      MPI_WIN_ALLOCATE_SHARED
> 
> On Mon, Sep 10, 2012 at 04:45:15PM -0500, Rajeev Thakur wrote:
>> I think the wording Jim suggested is the simplest for now:
>> 
>> If the window was created with \mpifunc{MPI\_WIN\_ALLOCATE} or
>> \mpifunc{MPI\_WIN\_ALLOCATE\_SHARED}, \mpifunc{MPI\_WIN\_FREE} will
>> free the window memory that was allocated during window creation.
> I'd really like to avoid such lists. I vote to add a sentence to the two
> creation functions that win_free (with a forward reference) will free
> the allocated memory (and remove the existing sentence in win_free).
> This will avoid inconsistencies in the future and seems rather
> uncontroversial.
> 
> Torsten
> 
> 
> --
> ### qreharg rug ebs fv crryF ------------- http://www.unixer.de/ -----
> Torsten Hoefler           | Assistant Professor
> Dept. of Computer Science | ETH Zürich
> Universitätsstrasse 6     | Zurich-8092, Switzerland
> CAB E 64.1                | Phone: +41 76 309 79 29
> _______________________________________________
> mpi3-rma mailing list
> mpi3-rma at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma
> 
> 
> Click https://www.mailcontrol.com/sr/kA9!XnKJn2vTndxI!oX7Uqfv98hA5qiofHZG1eUPj2QgAAvaR3Gz0LWZHFYheaKwErVAxTrKg1xGp10b+rkp2Q==  to report this email as spam.
> 
> _______________________________________________
> mpi3-rma mailing list
> mpi3-rma at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma