- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
I would like to run a heterogeneous program on a multi node cluster where each node comes with 2 MIC devices.
	Let's consider for example 2 dual-sockets nodes:
-------------------        -------------------
	--- NODE 1 --        --- NODE 2 --
	       |                            |
	 S    |  S                 S    |  S   
	 o     |  o                 o    |  o   
	 c     |  c                 c    |  c   
	 k     |  k                 k    |  k   
	 e     |  e                 e    |  e   
	 t      |  t                  t    |  t   
	 #1    |  #2              #1   |  #2  
	-------------------        -------------------
	MIC0  | MIC1         MIC0  | MIC1
	-------------------        -------------------
If I execute 4 mpi processes, how can I assure that each mpi process will run on a different socket, and that each mpi-rank will offload on the MIC device connected to the socket where it is executed?
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi Ale,
Did you try using "mpirun -ppn 2 -np 4 I_MPI_PIN_DOMAIN=socket ....".
Also I am assuming that you have Intel Xeon Phi coprocessors configured in external bridge topology and are able to ping each other across cluster.
Thanks
 
					
				
				
			
		
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page