Thanks for your reply. I am, indeed, specifically interested in the one-sided part :-) I'll send a mail to the MPI forum. Best, Matthieu ----- Mail original -----

De: "Jeff Hammond" <[email protected]> À: [email protected] Envoyé: Mercredi 24 Juillet 2013 09:16:56 Objet: Re: [mpich-discuss] availability of an implementation of MPI 3 on production platforms

If you post this to the MPI Forum list, Rich Graham should be able to share his slide deck of the ETA for MPI-3 for essentially all implementations.

MPICH-3 supports all of MPI-3 as far as I know. I think MVAPICH2-1.9 is derived from it but I have not verified that everything in MPI-3 is working (I have no reason to doubt this).

Word on the street is that Cray deployed their latest MPI implementation based upon MPICH-3, hence it should support MPI-3 as well.

The MPICH test suite exercises many aspects of MPI-3 and can be used to verify other implementations.

In the event you're interested in RMA, I have been using most of MPI-3 RMA from MPICH-3 regularly.

Best,

Jeff

On Wed, Jul 24, 2013 at 9:59 AM, Matthieu Dorier <[email protected]> wrote:

...

Hi,

This is a non-technical question:

The MPI 3 standard includes some very nice features, which I'm impatient to leverage in my projects, but being still recent, most platforms today (at least all production platforms I have access to) do not provide an implementation of it. I'm curious when do you think we will be able to say that a majority of platforms provide MPI 3, and that a code that using MPI 3 can be seen as portable? (how long did it take between the release of the MPI 2 standard and its availability on most platforms?) Are you aware of platforms that already provide it? And what is the status of the implementation in Mpich? (are there functions still to be implemented?)

Thanks

Matthieu Dorier PhD student at ENS Cachan Brittany and IRISA http://people.irisa.fr/Matthieu.Dorier

_______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss

-- Jeff Hammond [email protected] _______________________________________________ discuss mailing list [email protected] To manage subscription options or unsubscribe: https://lists.mpich.org/mailman/listinfo/discuss