On Fri, Dec 08, 2006 at 04:11:04PM +0530, krishna chaitanya wrote:
> I learnt from a reliable source that MPI uses the services
> provided by TCP/IP or infiniband.
Did you hear that from a whistle-blower?
> Suppose that there is a bottle-neck in
> the TCP/IP layer itself, how will optimization of MPI calls really help?
Bottleneck is usually a fuzzy term.
Collective calls basically means for example that you tell MPI:
"send this to all nodes"
for(i=0;i<number_of_nodes;i++) "send this to node i"
Therefore you don't have serialization of the tasks and the MPI layer
could do optmizations.
For example the connection network could have a way to broadcast or at
least sending could be interleaved.