[MPI3 Fortran] What if Fortran is the primary interface?

Aleksandar Donev donev1 at llnl.gov
Tue Mar 18 12:06:20 CDT 2008


On Tuesday 18 March 2008 07:26, Craig Rasmussen wrote:
> My test so far indicate that this works fine too.  As I recall,
>    type MPI_HANDLE
>        private
>        type(C_INT) :: handle
>    end type MPI_HANDLE
> worked for those implementations.  Since MPI_HANDLE is an opaque  
> type, the implementer has freedom to define it as she wishes.
To recap: This is likely to work on a lot of implementations (includes MPI+C 
ABI), but it is not guaranteed by the Fortran (or C) standards. C structs are 
different from plain integers and *may* be passed differently. Ultimately, I 
think if the design is such that it works fast on most (existing) 
implementations and *can* be implemented by wrappers on some more peculiar 
systems, it is OK.
Best,
Aleks

-- 
Aleksandar Donev, Ph.D.
Lawrence Postdoctoral Fellow @ Lawrence Livermore National Laboratory
High Performance Computational Materials Science and Chemistry
E-mail: donev1 at llnl.gov
Phone: (925) 424-6816  Fax: (925) 423-0785
Address: P.O.Box 808, L-367, Livermore, CA 94551-9900
Web: http://cherrypit.princeton.edu/donev




More information about the mpiwg-fortran mailing list