[MPI3 Fortran] MPI-2.1: Fortran 90 bindings

[Craig Rasmussen]

It is my hope that we can get a formal request from the MPI Forum for  
something like this to support MPI, although it might be better to  
change the restrictions on CLOC to do this.  The topic has come up a  
couple of times on the Fortran committee and failed for lack of  
support.  Coming from the MPI Forum (rather than just the few of us  
on the committee who use MPI) should help our cause.

(Aleks, you may want to join this email group, I know you have lots  
of free time :-)

Cheers,
Craig


On Mar 10, 2008, at 8:20 AM, Lionel, Steve wrote:

> Harald Klimach wrote:
>
>> It would be very nice, if Fortran would allow the assumed
>> size notation in a generic interface to consume all ranks.
>> Which would mean that you'd have to use assumed shape
>> arrays if you want to explicitly use different routines
>> for different ranks and both options would be mutually
>> exclusive.
>
> I discussed this with our Fortran standards rep last week. Intel and
> many other vendors have extensions that allow for this, but we agreed
> that the standard should be amended to allow for the option.  I
> suggested that as part of the C interoperability that a "VOID"  
> argument
> attribute be added which would serve the same purpose it does in C.
>
> Absent that, Fortran allows you to pass an element of an array of any
> rank to an array argument of any rank, such as:
>
> Call sub(a(1))
>
> Fortran specifies that the remaining elements in the actual argument
> array are "sequence associated" with the corresponding elements of the
> dummy argument.
>
> Fortran does NOT allow passing a scalar variable or expression to an
> array dummy argument.
>
> Steve Lionel
> Intel
>
> _______________________________________________
> mpi3-fortran mailing list
> mpi3-fortran at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-fortran