[MPI3 Fortran] Results of recent J3 Fortran meeting
longb at cray.com
Thu Oct 21 17:07:20 CDT 2010
N.M. Maclaren wrote:
> On Oct 21 2010, Bill Long wrote:
>> We currently interoperate only with C's char type. On the Fortran side,
>> the interoperable declaration is character(kind=c_char). Fortran's
>> default character kind is irrelevant. And sizeof(char) is 1. So there
>> is no issue here.
> Except for the last sentence, that is true. But one of MPI's major
> requirements is that it should be able to handle Fortran default types,
> and NOT force every program to use the C-interoperability KINDs to
> call MPI! Sorry, but there IS an issue.
OK, I agree that there is a general issue for MPI with non-interoperable
types. We have a type code for that - CFI_type_unspecified. I assume
that, if the MPI routines are expecting data objects of Fortran types
that have no corresponding C types (and hence there is little that can
be done with them in C other than memory copies) then the actual intent
is to only make memory copies. In that case, knowing the size of a
single character is unimportant. All you need is the base_addr,
elem_len, rank, and dim information.
The origin of CFI_type_unspecified is to provide a type code in the case
that the actual argument is TYPE(*). If you want to pass a
non-interoperable kind of character to a C function, you can write the
Fortran interface with a TYPE(*) argument. If the rest of the
declaration requires a C descriptor pointer in the C prototype, then
the contents of that descriptor will look just like the one I'm
proposing for the interoperable character kind, except that the type
code will be CFI_type_unspecified, rather than CFI_type_char. The
base_addr, elem_len, rank, and dim information will be correct and the
MPI memory copies should work just fine.
In fact, consistency in the way character arguments are passed to
CHARACTER(kind-C_CHAR) dummies and to TYPE(*) dummies is yet another
argument against the rank modification scheme.
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