Can you pad data so you can use MPI_Gather instead? It's possible that
Gatherv doesn't use recursive doubling.
Or you can implement your own aggregation tree to work around the in-cast
problem.
Jeff
On Thu, Sep 21, 2017 at 2:01 PM saiyedul islam
wrote:
> Hi all,
>
> I am working on paralleliza
Thank you very much for the reply, Sir.
Yes, I can observe this pattern by eyeballing Network communication graph
of my cluster (through Ganglia Cluster Monitor :
http://ganglia.sourceforge.net/).
During this loop execution, master is receiving data at ~ 100 MB/sec (of
theoretical 125 MB/sec in Gi
All your processes send their data to a single destination, in same time.
Clearly you are reaching the capacity of your network and your data
transfers will be bound by this. This is a physical constraint that you can
only overcome by adding network capacity to your cluster.
At the software level
Sorry for the noise. I found MPI_Init_thread and installed 1.2.4.
Seems to be fine now!
Thanks for the great work on the multi-threaded MPI codes!
Brian
On Nov 7, 2007 8:04 PM, Brian Budge wrote:
> Hi All -
>
> I am working on a networked cache for an out-of-core application, and
> currently
