Hi,
We are working right now on the new RMAPS component to allow rank mapping
according to the host : socket : core.
This component will be general and flexible. You will need to support the
rankmap file in the command line that looks similar to hostfile and the
component will do the rest.
I think the goal was to allow manual specification of this stuff
(E.g., in a text file) even though it could be fairly unwieldy in some
cases (specifying every core in every machine). Then use the
experiences learned from that to work on better syntax / specification
mechanisms.
On Feb
Hi;
Did openMPI tested on Cell BE?
Thank you in advance
I'm not entirely sure what you mean, but suspect the answer is probably
"no"...
On 2/6/08 8:48 AM, "Resat Umit Payli" wrote:
> Hi;
>
> Did openMPI tested on Cell BE?
>
> Thank you in advance
>
>
> ___
> devel mailing list
> de...@open-mpi.org
> ht
Hi;
I was talking about the IBM's Cell Broadband Engine.
Thank you
On Feb 6, 2008 10:56 AM, Ralph H Castain wrote:
> I'm not entirely sure what you mean, but suspect the answer is probably
> "no"...
>
>
> On 2/6/08 8:48 AM, "Resat Umit Payli" wrote:
>
> > Hi;
> >
> > Did openMPI tested on Ce
I know we have been using Open MPI with Cell processors in the past
(on PlayStations). However I don't know where is the status now, and I
think Open MPI was running on the attached power PC (and then data
transfered to the SPEs by the application).
Aurelien
Le 6 févr. 08 à 10:56, Ralph H
Hi Resat
I'm sorry I am not communicating clearly. I recognized the acronym. What I
don't fully understand is what you mean by "tested on Cell BE"?
If you mean "have we run Open MPI on a machine entirely composed of Cell
BE's", then the answer is "no". However, if you mean "have we run Open MPI
o
Hi Ralf;
Yes. There is a particular motivation. Here at IU last two day I did attend
the IBM workshop based on Cell programming. Right now I do have opportunity
to access this type of machine. I would like to play with my MPI based CFD
code. May be IBM will provide me some kind of MPI acceses on
At UTK we had a cluster of PS3. We wrote a documentation on how to use
the PS3 cluster, including how to configure and install Open MPI there
(there is a configuration file in the platform directory). You can use
the following link http://www.netlib.org/utk/people/JackDongarra/PAPERS/scop3.pd
Resat,
Right now there is no MPI implementation that will allow you to start
MPI processes on the SPEs. Keep this in mind when you try porting your
code.
george.
On Feb 6, 2008, at 1:06 PM, Resat Umit Payli wrote:
Hi Ralf;
Yes. There is a particular motivation. Here at IU last two day
Hi George;
Thank you for the link. I did see that link last night and dowload the
document already.
Thank you
2008/2/6 George Bosilca :
> At UTK we had a cluster of PS3. We wrote a documentation on how to use
> the PS3 cluster, including how to configure and install Open MPI there
> (there is a
Uwell, actually, yes there is...but not in release or in the OMPI
repository. It may become available sometime this summer.
On 2/6/08 11:17 AM, "George Bosilca" wrote:
> Resat,
>
> Right now there is no MPI implementation that will allow you to start
> MPI processes on the SPEs. Keep t
Hi folks,
WHAT:
Merge the htor-nbc branch to the trunk. This branch adds LibNBC support
to Open MPI (via compiler-wrappers mpi{cc,CC,xx,f90,..
Hi Ralph;
Whenever is ready I can test it.
Thank you
On Feb 6, 2008 2:57 PM, Ralph H Castain wrote:
> Uwell, actually, yes there is...but not in release or in the OMPI
> repository. It may become available sometime this summer.
>
>
> On 2/6/08 11:17 AM, "George Bosilca" wrote:
>
> > R
14 matches
Mail list logo
