[Mpi3-rma] MPI_Fetch_and_op and predefined datatypes

Jeff Hammond jhammond at alcf.anl.gov
Sat Oct 27 08:56:00 CDT 2012

Those types are listed in Section 5.9.4, not Section 5.9.2.  As
Hubert's original proposal made explicit reference to 5.9.2 and not
"anything predefined for reductions" (or equivalent), I think that his
text changes exclude the types that were designed to work with
MPI_MINLOC and MPI_MAXLOC.  Does F*Op work with those predefined
operations?  What does it even mean to do those in F&Op anyways?


On Fri, Oct 26, 2012 at 11:34 PM, Jim Dinan <dinan at mcs.anl.gov> wrote:
> Did we also intend to include pair types for F&Op?  They are allowed in
> reductions.
>  ~Jim.
> On 10/26/12 4:07 PM, Torsten Hoefler wrote:
>> On Fri, Oct 26, 2012 at 04:22:02PM -0500, Jeff Hammond wrote:
>>> I don't want to wait for 3.1.  Are we seriously going to require
>>> MPI_WCHAR support in F&Op to be MPI-3 compliant?  That makes no sense.
>> I think the intent of "Any of the predefined operations for MPI_REDUCE,
>> as well as MPI_NO_OP or MPI_REPLACE, can be specified as op;
>> user-defined functions cannot be used." is clear. One could even dream
>> up a use-case for MPI_WCHAR if one wants to emulate an atomic read
>> (using MPI_NO_OP).
>> I will not vote against an errata if you want to propose text and
>> initiate the process.
>> Best,
>>    Torsten
> _______________________________________________
> mpi3-rma mailing list
> mpi3-rma at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi3-rma

Jeff Hammond
Argonne Leadership Computing Facility
University of Chicago Computation Institute
jhammond at alcf.anl.gov / (630) 252-5381

More information about the mpiwg-rma mailing list