-------- Message d'origine --------
Sujet: Re: LAM: Re: Who's using MPI-2 features of LAM? (fwd)
Date: Mon, 14 Mar 2005 14:42:33 +0100
De: Sophie Valcke <Sophie.Valcke_at_[hidden]>
A: lam_at_[hidden]
CC: etienne.gondet_at_[hidden]
To who it may concern,
This is just to confirm that both versions of the Oasis coupler, OASIS3
(http://www.cerfacs.fr/globc/software/oasis/index.html) and
OASIS4 (http://www.cerfacs.fr/PRISM/WP3A/oasis4_UserGuide_No3_1.pdf)
use the MPI2 feature MPI_Comm_Spawn or MPI_Comm_spawn_multiple. OASIS is a
coupler of climate modelling codes used, in particular, in the framework
of the PRISM project (http://prism.enes.org/). This is very useful as it
allows the models being spawned to keep their own MPI_COMM_WORLD context.
Best regards,
Sophie Valcke, for the Oasis team.
---------------------------------------------------------------
Sophie Valcke
Permanent affiliation: Currently visiting at:
CERFACS - SUC URA 1875 Environnement Canada
42 Av. Coriolis Direction de la Recherche en Meteorologie
31057 Toulouse 2121 Route Transcanadienne, 5e etage
FRANCE Dorval (Quebec), H9P 1J3, CANADA
Tel: 05.61.19.30.76 Tel: 1-514-421-4678
Fax: 05.61.19.30.00 Fax: 1-514-421-2106
http://www.cerfacs.fr/globc/people/valcke.html
---------- Forwarded message ----------
Date: Wed, 02 Mar 2005 16:19:35 +0100
From: etienne gondet
To: General LAM/MPI mailing list , palmdvp_at_[hidden],
Sophie Valcke
Subject: Re: LAM: Re: Who's using MPI-2 features of LAM?
Hi,
There is the PALM project http://wwww.cerfacs.fr/~palm who use mpi2
especially
mpi_spawn the port mechanism but also mpi_put and get through the lock
unlock mechanism
using japanese mpi versions and most of the time the excellent lam
implementation of mpi 2.
The reserach version is based upon mpi1 the PALM_MP version used both
mpi1 + mpi2 and is already released.
The status is free under agreement for research project.
This a dynamic coupler for parallel codes especially used by several
projects especially in "climate oceanographic, ...
topics but also for few CFD coupling studies.
There is also a mpi2 version of the oasis coupler (still from
cerfacs) more specifically designed for coupling climatic code :
athmosphere, ocean, land, ice, biology, ...
Sincerely yours.
Etienne Gondet
HPC coordinator
GIP MERCATOR Océan
8-10 rue HERMES
31526 RAMONVILLE SAINT-AGNE CEDEX
France
tel: 05 61 39 38 11 (from France); 33 5 61 39 38 11 (other)
fax: 05 61 39 38 99 (from france); 33 5 61 39 38 99 (other)
Bryan O'Sullivan a écrit:
>On Mon, 2005-02-07 at 15:28 -0800, Bryan O'Sullivan wrote:
>
>
>
>>I can summarise responses for the list, if anyone is interested.
>>
>>
>
>The answer, for those who are interested, is that apparently nobody uses
>MPI-2 at all. One person responded that he'd have used non-blocking
>collectives if they had been available in an MPI implementation when he
>was writing his app. Another noted that she couldn't find any MPI-2
>code in the wild, either.
>
>
>_______________________________________________
>This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
>
>
>
|
