[MPI3 Fortran] MPI non-blocking transfer

N.M. Maclaren nmm1 at cam.ac.uk
Thu Feb 12 02:21:42 CST 2009


On Feb 11 2009, Aleksandar Donev wrote:
>On Wednesday 11 February 2009 09:08, Craig Rasmussen wrote:
>
>> Ignoring the precise syntax for now, the important point is whether
>> an   attribute on a procedure will be sufficient.
>
>And that question has been answered with "no" in discussions: the data 
>(buffer) itself must be marked somehow, not just the Send/Recv/Wait 
>call site. ...

Yes.  An attribute of the transfer buffer is both necessary and
sufficient.  A property of a procedure is neither sufficient nor
even helpful.  That is PRECISELY why the ASYNCHRONOUS attribute is
an attribute of objects!


Regards,
Nick Maclaren,
University of Cambridge Computing Service,
New Museums Site, Pembroke Street, Cambridge CB2 3QH, England.
Email:  nmm1 at cam.ac.uk
Tel.:  +44 1223 334761    Fax:  +44 1223 334679





More information about the mpiwg-fortran mailing list