On Mar 1, 2016, at 2:28 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote:
>
>
> On Tue, Mar 1, 2016 at 2:12 PM, Bill Long <longb(a)cray.com> wrote:
>
> On Mar 1, 2016, at 2:03 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote:
>
> > https://github.com/mpi-forum/mpi-issues/issues/43
> >
> > any objections?
>
> Not from me. It seems like an excellent idea.
>
>
> The only question is if MPI_Type_create_darray makes sense to be used with Fortran coarrays for MPI file I/O, but this is not something I can reason about, because I do not understand MPI I/O. Also, it is not clear to me that MPI_Type_create_darray is really what people want to use, given we have MPI_Type_subarray.
I’m even less of an expert about MPI I/O. However, it would seem very unlikely that a coarray would be treated any different from an ordinary array. A coarray is just an ordinary array with the extra feature of being accessible from a different image. On the local image, it is just a ordinary array. Since I really doubt that MPI I/O is using coarray features inside, it seems unlikely that coarrays would be special in any way for MPI I/O. For defining coarray elements by reading a file, I would not use MPI I/O at all. Just have each image read in its part of the file - time tested solution. For writing, the array is memory is not modified, so even if MPI I/O is used, I would not see an issue of the output item were a coarray.
Cheers,
Bill
>
> Jeff
>
> Cheers,
> Bill
>
>
> >
> > jeff
> >
> > --
> > Jeff Hammond
> > jeff.science(a)gmail.com
> > http://jeffhammond.github.io/
> > _______________________________________________
> > mpiwg-fortran mailing list
> > mpiwg-fortran(a)lists.mpi-forum.org
> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran
>
> Bill Long longb(a)cray.com
> Fortran Technical Support & voice: 651-605-9024
> Bioinformatics Software Development fax: 651-605-9142
> Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101
>
>
> _______________________________________________
> mpiwg-fortran mailing list
> mpiwg-fortran(a)lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran
>
>
>
> --
> Jeff Hammond
> jeff.science(a)gmail.com
> http://jeffhammond.github.io/
> _______________________________________________
> mpiwg-fortran mailing list
> mpiwg-fortran(a)lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran
Bill Long longb(a)cray.com
Fortran Technical Support & voice: 651-605-9024
Bioinformatics Software Development fax: 651-605-9142
Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101
On Tue, Mar 1, 2016 at 2:12 PM, Bill Long <longb(a)cray.com> wrote:
>
> On Mar 1, 2016, at 2:03 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote:
>
> > https://github.com/mpi-forum/mpi-issues/issues/43
> >
> > any objections?
>
> Not from me. It seems like an excellent idea.
>
>
The only question is if MPI_Type_create_darray makes sense to be used with
Fortran coarrays for MPI file I/O, but this is not something I can reason
about, because I do not understand MPI I/O. Also, it is not clear to me
that MPI_Type_create_darray is really what people want to use, given we
have MPI_Type_subarray.
Jeff
> Cheers,
> Bill
>
>
> >
> > jeff
> >
> > --
> > Jeff Hammond
> > jeff.science(a)gmail.com
> > http://jeffhammond.github.io/
> > _______________________________________________
> > mpiwg-fortran mailing list
> > mpiwg-fortran(a)lists.mpi-forum.org
> > http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran
>
> Bill Long
> longb(a)cray.com
> Fortran Technical Support & voice:
> 651-605-9024
> Bioinformatics Software Development fax: 651-605-9142
> Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101
>
>
> _______________________________________________
> mpiwg-fortran mailing list
> mpiwg-fortran(a)lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran
>
--
Jeff Hammond
jeff.science(a)gmail.com
http://jeffhammond.github.io/
On Mar 1, 2016, at 2:03 PM, Jeff Hammond <jeff.science(a)gmail.com> wrote:
> https://github.com/mpi-forum/mpi-issues/issues/43
>
> any objections?
Not from me. It seems like an excellent idea.
Cheers,
Bill
>
> jeff
>
> --
> Jeff Hammond
> jeff.science(a)gmail.com
> http://jeffhammond.github.io/
> _______________________________________________
> mpiwg-fortran mailing list
> mpiwg-fortran(a)lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-fortran
Bill Long longb(a)cray.com
Fortran Technical Support & voice: 651-605-9024
Bioinformatics Software Development fax: 651-605-9142
Cray Inc./ Cray Plaza, Suite 210/ 380 Jackson St./ St. Paul, MN 55101
