On Wed, 2 Nov 2005, Carsten Kutzner wrote:
> In our case the congestion happened within the switch.
Can you explain how you came to this conclusion (with as many details
as possible) ?
> I therefore experimented with the all-to-all code and have now a
> modified Alltoall for multi-CPU nodes which shows no congestion even
> when Ethernet flow control is turned off.
We also had a bit of off-list discussion about this and we came up
with 2 algorithms for avoiding congestion. Buf Jeff didn't say
anything afterwards regarding a possible inclusion, probably due to
Open MPI taking all his time ;-)
> The performance of the original LAM MPI_Alltoall however remains a
> bit better for small message sizes. This is similar to what Pierre
> found for his modified routines.
Then you can try to use the original for small messages and the new
one for large messages, with some threshold value to switch from one
to the other.
Due to the common use of the same switch by you and Pierre, maybe it's
possible to find some "optimized" conditions for this particular piece
of hardware...
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: Bogdan.Costescu_at_[hidden]
|
