Newsgroups: comp.parallel.mpi
From: gropp@godzilla.mcs.anl.gov (William Gropp)
Subject: Re: reductions deterministic?
Organization: MCS, Argonne National Laboratory
Date: Sat, 16 Mar 1996 15:8:47 GMT
Message-ID: <8269889271688@godzilla.mcs.anl.gov>

In article <4ics0k$3e8@newshost.lanl.gov>,
Richard Barrett <rbarrett@ranger.lanl.gov> wrote:
>In its "advice to implementors" it is "strongly recommended" that MPI reductions
>be deterministic. Can anyone please tell me if this recommendation has
>been followed by:
>
>- MPICH (including MPP versions) 

Yes.  But note that deterministic applies to the exact same situation (number
of processors, order of members in group).  It says nothing about differing
numbers of processors (which some would also like, but which seriously 
impacts performance).

Bill

