3 Apr
2011
3 Apr
'11
6:14 p.m.
On 03/31/2011 10:20 PM, Rajeev Thakur wrote:
Good catch. I think we should add that restriction. MPI_Sendrecv, for example, explicitly says the send and recv buffers must be disjoint.
No MPI_IN_PLACE for Get_accumulate? Not yet, I agree that it could make sense though -- we should discuss this tomorrow.
Could you perhaps provide a use case? Otherwise, it looks like an implementation nightmare with potentially unexpected performance characteristics associated with any working implementation choice. Do we have a call tomorrow? I thought Rajeev's note said 4/11...