George,

Thank you for your answer. Is [1] the paper mentioned by Jeff in an earlier 
message (ICS 2004)?
Thanks

Matthieu
________________________________
From: users [[email protected]] on behalf of George Bosilca 
[[email protected]]
Sent: Tuesday, April 19, 2016 2:03 PM
To: Open MPI Users
Subject: Re: [OMPI users] MPI_Bcast implementations in OpenMPI

Matthieu,

If you are talking about how Open MPI selects between different broadcast 
algorithms you might want to read [1]. We have implemented a dozen different 
broadcast algorithms and have run a set of tests to measure their performance. 
We then used a quad tree clasiffication algorithm to minimize the number of 
switch points, and then generated the C code to dynamically make the selection 
based on the message size and the number of participants. Take a look at the 
function ompi_coll_tuned_bcast_intra_dec_fixed in 
ompi/mca/coll/tuned/coll_tuned_decision_fixed.c.

 George.



On Tue, Apr 19, 2016 at 4:13 PM, Jeff Squyres (jsquyres) 
<[email protected]<mailto:[email protected]>> wrote:
On Apr 15, 2016, at 9:18 AM, Dorier, Matthieu 
<[email protected]<mailto:[email protected]>> wrote:
>
> I'd like to know how OpenMPI implements MPI_Bcast. And if different 
> implementations are provided, how one is selected.

This is a fairly complicated topic.  This old paper is the foundation for how 
Open MPI works (it's a bit different these days, but the basic concepts 
described in this paper are still mostly relevant):

    https://www.open-mpi.org/papers/ics-2004/

--
Jeff Squyres
[email protected]<mailto:[email protected]>
For corporate legal information go to: 
http://www.cisco.com/web/about/doing_business/legal/cri/

_______________________________________________
users mailing list
[email protected]<mailto:[email protected]>
Subscription: http://www.open-mpi.org/mailman/listinfo.cgi/users
Link to this post: 
http://www.open-mpi.org/community/lists/users/2016/04/28967.php

Reply via email to