On 5/7/13 5:57 PM, Damon Catling wrote:
> I have no idea weather Nvidia and Valve already have their own remote 3D
> solutions for Linux similar to VirtualGL, but if not, perhaps VirtualGL
> would be attractive to companies of that ilk?

Valve, possibly.  nVidia has their own solution called VGX, which is 
basically a mechanism for GPU sharing among virtual machines.  It's 
proprietary and vendor-locked (you have to buy specific hardware to use 
it, and you have to buy and renew client licenses for every machine that 
connects to it.)  If anyone had any doubts that nVidia is the new SGI ...


> Valve have been sleuthing for talent recently:
> http://www.valvesoftware.com/jobs/job_postings.html

Any company dealing in the video game market at all is going to be 
looking for people who are typically more passionate about video games 
than I am (I think the last one I actually played actively was probably 
Doom, and that was at least in part because I was getting paid to test 
it.)  They rely on that passion when they ask you to work until 2 AM to 
get a release out the door.  It's a game for the young.


> The other thing that comes to mind is that sooner or later Linux distros
> will be moving over to Wayland, and Mir on Ubuntu. Is VirtualGL likely
> to slide smoothly into that change, or could it be a major hurdle for
> the continuation of VirtualGL?

Well, yes and no.  From the point of view of an X app, Wayland is just 
another layer that sits underneath the X server, so as long as the 
necessary GLX interfaces are in place and are still accelerated, then 
VirtualGL can continue to work for X apps as it currently does. 
However, for new apps that are no longer written as X apps, then we'd 
need a new interposer.  My limited understanding of Wayland is that it 
will provide a kind of equivalent to GLX, a set of EGL functions that 
allow OpenGL contexts to be established in specific windows.  If so, 
then conceptually, the new interposer would work like the current one-- 
intercepting the context functions, redirecting the contexts into 
off-screen buffers, reading back the frame buffer at the appropriate 
time, and compositing it into the window.  The big question marks are: 
How does a typical Wayland app draw an image?  How does Wayland work for 
remote display?  Does it retain the concept of different "servers"?  For 
instance, can I be running a Wayland server connected to the hardware, 
then another one that renders into a virtual framebuffer (a la VNC)?  Do 
OpenGL apps under Wayland already render off-screen?

Without having some of those specifics nailed down, it would be 
impossible to say how VirtualGL should be implemented under this new 
architecture.  Perhaps an interposer isn't even necessary.

------------------------------------------------------------------------------
Learn Graph Databases - Download FREE O'Reilly Book
"Graph Databases" is the definitive new guide to graph databases and 
their applications. This 200-page book is written by three acclaimed 
leaders in the field. The early access version is available now. 
Download your free book today! http://p.sf.net/sfu/neotech_d2d_may
_______________________________________________
VirtualGL-Users mailing list
VirtualGL-Users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/virtualgl-users

Reply via email to