[mpiwg-rma] FW: [Mpi3-rma] [EXTERNAL] Re: MPI-3 UNIFIED model updates
Underwood, Keith D
keith.d.underwood at intel.com
Wed Aug 28 18:02:25 CDT 2013
> >>> Also, I think we are ignoring the point Jim raised: does OpenSHMEM
> >>> really need these semantics?
> >> I think this is a grey area of OpenSHMEM and my belief is that it
> >> really does need those semantics.
> >Perhaps this is just an oversight in OpenSHMEM and can be fixed?
> I don't think it's an oversight. Remember that OpenSHMEM, while a great
> leap forward from SGI's man pages, is still a fairly loose specification.
> There are definitely codes that spin on a completion word (or poll and move
> on to other things, where shmem_wait wouldn't work) in the wild, and
> they're considered correct codes.
And, I think they *should* be correct. It should be possible in OpenSHMEM and MPI3-RMA-UNIFIED to do shared memory programming using threads locally in the symmetric heap/unified window and have the same mechanisms "just work" when updates from remote accesses arrive. This provides for an "easy" to access programming model while providing explicit remote memory accesses.
More information about the mpiwg-rma