So in this case, pvserver is using the hardware GPU to render to an off-screen context? It must also then be doing a pixel readback before it sends the image to the client? On my server I have several proceesors and one gfx card. If I do something like: mpirun -np 4 pvserver --use-offscree-rendering will each process create an off-screen rendering conext using the same gfx card? In this case are the four images read back and combined before they are sent to the client? Thanks for the answers.
________________________________ From: Moreland, Kenneth [mailto:[email protected]] Sent: Thu 2/5/2009 2:27 PM To: Jim Montine; David E DeMarle Cc: [email protected] Subject: Re: [Paraview] remote rendering with paraview You need the DISPLAY environment variable because the standard way to get a gfx context on *nix is through an X server. To create an off-screen rendering context you first have to connect to an X server. It is also a convenient way to specify which GPU you want to use if a system has more than one. There is one exception to the DISPLAY/X server requirement. ParaView supports the use of OSMesa, which allows you to open a Mesa 3D rendering context without an X server. This simplifies running pvserver a great deal, but works exclusively with software only rendering. Thus, we recommend it if and only if you have no graphics hardware. -Ken On 2/5/09 1:26 PM, "Jim Montine" <[email protected]> wrote: In case 2 below (where the server optionally renders the polygons), why is it necessary to set the DISPLAY to the gfx head of the server (things are displayed on the PC) Why is the DISPLAY environment needed at all for this off-screen rendering case? **** Kenneth Moreland *** Sandia National Laboratories *********** *** *** *** email: [email protected] ** *** ** phone: (505) 844-8919 *** web: http://www.cs.unm.edu/~kmorel
_______________________________________________ ParaView mailing list [email protected] http://www.paraview.org/mailman/listinfo/paraview
