I am expecting to take delivery of a small cluster of nvidia GPU enable nodes (each fully capable as machine with keyboard and mouse etc)
I am wondering if there is a mode or way of using Paraview so that I can divide up a display on one machine and get the OpenGL rendering done on 4 of the machines?
Is this something which requires specialist knowledge or hardware ? Or is that fairly common and built into Paraview ?
Where can I start reading up on such parallel OpenGL/GPU usage ?
While not very common, it is documented and used.
Thank you. One point of clarification.
I plan to have 5 workstation, one acting as the head node which will have the main display (Full HD)
Reading IceT and tiled display, it refers to multiple monitor but in my setup, I have one display monitor.
Is IceT able to tile up a single Full HD into appropriate chunk and have the 4 workstation GPU render and composite them back seamlessly into a single Full HD image for user interaction on the head node?
the comand “mpirun -bynode -np 8 ./pvserver -display :0.0 : -np 8 ./pvserver -display :0.1” is already out of date?
Now the option is “–displays”: like pvserver --displays=0,1?
I am testing Nvidia Index seems even “–displays” still not work for IndeX?
I will create a new thread to ask this question.
Thanks and Best Regards