[Mpi3-rma] missing RMA info keys

Jim Dinan dinan at mcs.anl.gov
Mon Dec 3 10:37:58 CST 2012


On 12/3/12 10:33 AM, Jeff Hammond wrote:
> yes, to summarize what Hubert and Jim said and add my own comments:
>
> * accumulate_ordering and accumulate_ops should be valid for all
> window types (unless shared memory windows allow ONLY load-store, then
> we can remove them from MPI_WIN_ALLOCATE_SHARED).

Shared memory windows are not restricted to only load/store accesses. 
This info key should be valid for all window types.

> * same_size is primarily useful for MPI_WIN_ALLOCATE but it would be
> valid and mildly useful for MPI_WIN_CREATE.  for example, if i want to
> do valid-range checking and i don't know same_size, i may need an O(N)
> list of (base,max_offset), whereas same_size allows me to have only
> O(N) for base and O(1)=1 for max_offset.

same_size is potentially useful for MPI_Win_allocate_shared, and is not 
useful for MPI_Win_create_dynamic.

> * alloc_shared_noncontig only makes sense for MPI_WIN_ALLOCATE_SHARED

Yep.

  ~Jim.



More information about the mpiwg-rma mailing list