[Paraview] Running pvserver 4.1 with 2 GPUs on a Single Node

[D H]

Hi,

I am working with some large VTK files (4GB each), and I'm attempting to setup a Paraview client/server to render them.  I'm working with the latest builds of both the client and server (4.1.0-277-g1bbd3c6), with the server on Ubuntu 12.04 and the client on Mac OS X 10.9.

I've been reading the documentation on the Paraview wiki about setting up the server to run on multiple processors and GPUs, but I'm still not sure of the exact syntax I should be using when I run pvserver.  I have one node with 16 cores and 2 NVIDIA Tesla GPUs, and I'd like to use all these resources for processing/rendering.

I can run commands like "mpirun -np 16 pvserver", and I'm successfully able to launch and connect to the server, but I don't know whether pvserver is then using both of my GPUs.  http://paraview.org/Wiki/Setting_up_a_ParaView_Server refers to doing something like "mpirun -bynode -np 8 ./pvserver -display :0.0 : -np 8 ./pvserver -display :0.1" but setting the environment variable DISPLAY instead of using -display.  But I'm not sure how I would find out what values to set for DISPLAY.  First, I'm not sure where my displays are located - is there any way to (a) test that localhost:0.0 and localhost:0.1 are correct, or (b) if those aren't correct, find our which ones are?  Second, what is the correct syntax for setting DISPLAY to have multiple values?  Something like 'export DISPLAY="localhost:0.0;localhost:0.1"'?

Assuming I have DISPLAY set to the correct value, is the correct syntax for launching pvserver under my configuration "mpirun -np 8 pvserver : -np 8 pvserver"?

I really appreciate your assistance in figuring this out.  Thank you very much for your help!

Best regards,

David