[MPI3 Fortran] Results of San Jose Forum meeting
longb at cray.com
Thu Mar 11 11:43:27 CST 2010
Jeff Squyres wrote:
> On Mar 11, 2010, at 8:59 AM, N.M. Maclaren wrote:
>>> That being said, this is definitely an engineering tradeoff. As I
>>> mentioned, the current OMPI prototype implementation has a Fortran
>>> derived type that includes the corresponding OMPI C handle as a member
>>> (which is a pointer). I was a bit unhappy to give this up, but Rolf
>>> convinced me that the other benefits outweigh this.
>> Does it still hold true, if it doesn't do what he though that it did?
> (I'm stepping back and letting you Fortran guys debate all the other points)
> There certainly is an elegance to having no MPI handle conversion functions between use mpi/use mpi3/mpif.h. It would have been nice to avoid the lookup, but a) the cost really doesn't matter, and b) I think the usability issue is nice (i.e., assignment), even if none of the other benefits work out. But I'm not tied strongly to either way.
One advantage of the integer flag values is that it avoids interface
issues when we change the pointer size. This happened many years ago
when we switched from 32 bits to 64 bits for addresses (and in truly
ancient history from 16 to 32), and the 128-bit address train is rapidly
Bill Long longb at 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
More information about the mpiwg-fortran