[Mpi3-rma] [EXTERNAL] Re: request-based ops
balaji at mcs.anl.gov
Sat Jun 15 20:44:03 CDT 2013
On 06/15/2013 12:42 PM, Jeff Hammond wrote:
> Related to this, I was wondering about the semantics of IWIN_FLUSH
> when ordering was disabled for a window (i.e. accumulate_ordering="").
> If I do the following, can I make any assumptions about the ordering
> of the two accumulates?
> MPI_Win win_with_unordered_accumulates;
> MPI_Request req1, req2;
> The Waitall is not required to act on the requests in order, right?
> Thus is req2 is completed first, then it flushes both Accumulates
> simultaneously, hence does not ordering them. The completion of req1
> is then trivial and have no effect on ordering Accumulate 1 before
> Accumulate 2.
Correct. No additional ordering constraints of completion will be added
to WAITALL. Like with any requests, they can complete in whatever order.
More information about the mpiwg-rma