[ofiwg] proposal to add a new completion semantic

[Kenneth Raffenetti]

On 08/02/2018 05:57 PM, Hefty, Sean wrote:
>> MPI synchronous send does not guarantee data placement into the target
>> buffer. It only guarantees that the recv operation on the target side
>> has been posted and matched. The data can still come later.
> 
> How about this change then?
> 
>>> FI_MATCH_COMPLETE*
>>> : Indicates that a completion should be generated only after the
>> operation
>>>     has been placed into the application specified buffer.
> 
>                  ^^^^^^^^^^^^^^^
>                  matched with an

Looks good to me.

> 
>> Operations using
>>>     this completion semantic are dependent on the application at the
>> target
>>>     claiming the message or results.  As a result, match complete may
>> involve
>>>     additional provider level acknowledgements or lengthy delays.
>> However, this
>>>     completion model enables peer applications to synchronize their
>> execution.