[mpiwg-rma] RMA Errata

Jeff Hammond jeff.science at gmail.com
Wed Jan 28 12:46:56 CST 2015


Fine, but it would be helpful if you would respond to my comments
about allowing point-to-point _synchronization_ on shared memory using
PSCW, since those were the more germane ones anyways.

Thanks,

Jeff

On Tue, Jan 27, 2015 at 11:12 PM, William Gropp <wgropp at illinois.edu> wrote:
> No, this routine only tells you what address a particular process (specified
> by rank) uses as the base address of a particular segment of shared memory.
> It doesn’t say anything about which process “owns” the shared memory or is
> “associated” with it.
>
> The issue is really one of care and precision in terminology, as well as
> consistency with the language and definitions used by others in describing
> shared memory.  MPI shared memory RMA will not be credible if we make up own
> own vague terms.
>
> Bill
>
> On Jan 28, 2015, at 6:45 AM, Jeff Hammond <jeff.science at gmail.com> wrote:
>
> And it is absolutely clear in MPI-3 that shared memory is associated
> with a particular process.  See the rank argument in
> http://www.mpich.org/static/docs/v3.1/www3/MPI_Win_shared_query.html.
> This function takes process rank as an _input_, which means that there
> is a well-defined notion of memory associated with a particular
> process.
>
>
>
> _______________________________________________
> mpiwg-rma mailing list
> mpiwg-rma at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-rma



-- 
Jeff Hammond
jeff.science at gmail.com
http://jeffhammond.github.io/



More information about the mpiwg-rma mailing list