[mpiwg-hybridpm] Hybrid telecon fiasco

[Jeff Squyres (jsquyres)]

Ok.  

I think we need to be very, very careful with terminology here -- there are *not* multiple MPI processes.  There are multiple endpoints in a single MPI process -- big difference (especially in terms of threaded resources in the same process).



On Apr 7, 2014, at 6:16 PM, Jim Dinan <james.dinan at gmail.com> wrote:

> That's my thinking -- we do say that each endpoint (i.e. communicator handle that is returned) behaves as an MPI process.
> 
> 
> On Mon, Apr 7, 2014 at 4:25 PM, Jeff Squyres (jsquyres) <jsquyres at cisco.com> wrote:
> On Apr 7, 2014, at 2:07 PM, Jim Dinan <james.dinan at gmail.com> wrote:
> 
> > Every process in a communicator has its own set of attributes.
> 
> "Every process in a communicator" -- does this mean each communicator handle that is returned?
> 
> --
> Jeff Squyres
> jsquyres at cisco.com
> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
> 
> _______________________________________________
> mpiwg-hybridpm mailing list
> mpiwg-hybridpm at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-hybridpm
> 
> _______________________________________________
> mpiwg-hybridpm mailing list
> mpiwg-hybridpm at lists.mpi-forum.org
> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpiwg-hybridpm


-- 
Jeff Squyres
jsquyres at cisco.com
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/