[Mpi-forum] MPI user survey

[Barrett, Brian W]

Alexander -

Who wouldn't answer such a question with "yes"?  I think such a question is
only meaningful if we know what users have to give up and what they get in
return.  For example:

I [strongly agree...strongly disagree] that I would give up the use of
communicators and datatypes in my application in return for a x% performance
improvement.

Otherwise, we're asking a question that really doesn't tell us anything we
don't already know.  Since I'm pretty positive we don't have the data to ask
such a question, it seems like waiting until next year when we might have a
better feeling for subsetting impact might be a good idea.

Brian


On 11/16/09 7:28 , "Supalov, Alexander" <alexander.supalov at intel.com> wrote:

> Hi,
> 
> Why don't we ask directly: do you want to have subsets in MPI-3 to trade
> feature richness for performance if you care?
> 
> Best regards.
> 
> Alexander
> 
> -----Original Message-----
> From: mpi-forum-bounces at lists.mpi-forum.org
> [mailto:mpi-forum-bounces at lists.mpi-forum.org] On Behalf Of Jeff Hammond
> Sent: Monday, November 16, 2009 4:23 PM
> To: Main MPI Forum mailing list
> Subject: Re: [Mpi-forum] MPI user survey
> 
> Based upon the issues that came up most of the meeting, we could
> append the question to:
> 
> x. MPI one-sided communication performance is more important to me
> than supporting a rich remote memory access (RMA) feature set.
> Complex user-defined datatypes and arbitrary remote atomics are
> examples of features that ~may~ lead to reduced performance for some
> MPI implementations and/or hardware architectures.
> 
> However, any such enumeration will lead to a minefield of user assumptions.
> 
> Jeff
> 
> On Mon, Nov 16, 2009 at 7:15 AM, Richard Treumann <treumann at us.ibm.com> wrote:
>> MPI one-sided communication performance is more important to me than
>> supporting a rich remote memory access (RMA) feature set.
>> Having been involved in several email discussions and knowing what is being
>> pressed, I can guess the desired answer. We cannot legitimately make
>> decisions based on such an ambiguous question: What expectations of typical
>> MPI communication are included in "rich"? Is using any communicator except
>> MPI_COMM_WORLD part of "rich"? Is getting a non-MPI_SUCCESS return code for
>> an error part of rich?
> 
> --
> Jeff Hammond
> Argonne Leadership Computing Facility
> jhammond at mcs.anl.gov / (630) 252-5381
> http://www.linkedin.com/in/jeffhammond
> http://home.uchicago.edu/~jhammond/
> 
> _______________________________________________
> mpi-forum mailing list
> mpi-forum at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
> ---------------------------------------------------------------------
> Intel GmbH
> Dornacher Strasse 1
> 85622 Feldkirchen/Muenchen Germany
> Sitz der Gesellschaft: Feldkirchen bei Muenchen
> Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer
> Registergericht: Muenchen HRB 47456 Ust.-IdNr.
> VAT Registration No.: DE129385895
> Citibank Frankfurt (BLZ 502 109 00) 600119052
> 
> This e-mail and any attachments may contain confidential material for
> the sole use of the intended recipient(s). Any review or distribution
> by others is strictly prohibited. If you are not the intended
> recipient, please contact the sender and delete all copies.
> 
> 
> _______________________________________________
> mpi-forum mailing list
> mpi-forum at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum
> 

--
   Brian W. Barrett
   Dept. 1423: Scalable System Software
   Sandia National Laboratories