On Dec 5, 2013, at 4:04 AM, ºúÑî <781578278_at_[hidden]> wrote:
> My ROCKS cluster includes one frontend and two compute nodes.In my program,I have use the openmpi API such as MPI_Send and MPI_Recv . but when I run the progam with 3 processors . one processor send a message ,other receive message .here are some code.
> MPI_Send(a,number, MPI_INT, 1, 1,MPI_COMM_WORLD);
> MPI_Recv(b, number, MPI_INT, 0, MPI_ANY_TAG, MPI_COMM_WORLD, &status);
> when number is least than 10000,it runs fast.
> but number is more than 15000,it runs slowly
Can you precisely define "fast" and "slowly"?
What is the network that MPI is using -- TCP over 1Gb ethernet?
I'm guessing that Open MPI is changing protocols (from an eager send to a rendezvous send) between these two sizes, but without more information, it's hard to say.
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/