[mpich-discuss] //Ssend(...) safe?

[Sören Freudiger]

Thanks @all for helping me out!

-----Ursprüngliche Nachricht-----
Von: mpich-discuss-bounces at mcs.anl.gov [mailto:mpich-discuss-bounces at mcs.anl.gov] Im Auftrag von Dorian Krause
Gesendet: Montag, 17. November 2008 11:04
An: mpich-discuss at mcs.anl.gov
Betreff: Re: [mpich-discuss] //Ssend(...) safe?

Sören Freudiger wrote:
>>> PS: I'm using Ssend, because possibly my send-call happens before my Irecv has been called. That means Rsend results in undefined behaivoir and that means standard Send shouldn't been used because reday send could be implemented for that one.
>>>   
>>>       
>
>   
>> I don't think so. As far as I know, MPI_Send may choose between 
>> MPI_Bsend and MPI_Ssend. Am I wrong?
>>     
>
> Hmm, mpi21-report.pdf - S.41 
>
> "... Ready send can be implemented as a standard send. In this case there will be no
> performance advantage (or disadvantage) for the use of ready send. ..."
>   

But that means that the implementors might just set MPI_Rsend = 
MPI_Send. In this case MPI_Rsend can be used without a posted receive 
(of course it is unsafe to assume this). MPI_Send is safe anyway.

> In my eyes that means that the implementation can use "ready send" for "standard send", can't it?
>   

The opposite way around.

Dorian

> SirAnn