[mpich-discuss] A quick question on collective routines - never mind

[Hiatt, Dave M]

a brain malfunction on my part, I need to have a bottle of wine, I got it, just been looking at the same problem too long!

-----Original Message-----
From: mpich-discuss-bounces at mcs.anl.gov
[mailto:mpich-discuss-bounces at mcs.anl.gov]On Behalf Of Hiatt, Dave M
[CH-REL]
Sent: Friday, January 08, 2010 4:36 PM
To: mpich-discuss at mcs.anl.gov
Subject: [mpich-discuss] A quick question on collective routines


I need to know if I've made a really dumb mistake.  The implications are big;

On a collective routine, like MPI:: Bcast, when it is broadcast out, do the compute nodes that read in the data have to always respond in the exact order (the ordering being rank) each time.  That is, if the first message is read by nodes 1,3,5,7,9,2,4,6,8,10 in that order, to all subsequent broadcasts on the same communicator for the lives of those processes have to occur in the same order.

I'm trying to make sense of the following sentence is "Using MPI".  "MPI requires that collective routines on the same communicator be called by all processes in that communicator , in the same order".

Thanks a whole lot for help

dave


"Consequences, Schmonsequences, as long as I'm rich". - Daffy Duck
Dave Hiatt
Market Risk Systems Integration
CitiMortgage, Inc.
1000 Technology Dr.
Third Floor East, M.S. 55
O'Fallon, MO 63368-2240

Phone:  636-261-1408
Mobile: 314-452-9165
FAX:    636-261-1312
Email:     Dave.M.Hiatt at citigroup.com




_______________________________________________
mpich-discuss mailing list
mpich-discuss at mcs.anl.gov
https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss