All of LAM's SMP collective algorithms are based upon the MagPIe work.
Do a few scholar.google.com searches for Magpie and MPI and you should
turn up the work. This stuff is integrated into LAM using the
component architecture -- the best paper of which is actually on the
Open MPI web site, but is essentially the same in LAM/MPI:
http://www.open-mpi.org/papers/ics-2004/.
We also have some shared memory collective algorithms, but there
haven't been any papers published about them mainly because they use
fairly well-known / understood shmem collective techniques.
On Dec 15, 2004, at 2:51 PM, death rince wrote:
> Hi Folks,
>
> I am trying to study to see how smp aware mpi
> implementations vs non-smp aware implementation works.
> I would appreciate if you could point me to documents
> that discuss the implementation in lam-mpi and to
> benchmarks/results of studies done to prove the
> effectiveness.
>
> thank you
> regards
> shiva
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
> _______________________________________________
> This list is archived at http://www.lam-mpi.org/MailArchives/lam/
>
--
{+} Jeff Squyres
{+} jsquyres_at_[hidden]
{+} http://www.lam-mpi.org/
|
