For large multi-screen setups, we've had success with matrox's triplehead2go product. We are running a setup that has a dual core machine with one 7900GTX and two TripleHead2Go boxes. As far as OSG is concerned its only rendering one 3840x2048 window with one camera on one rendersurface, which in reality is spread across 6 screens. We are rendering a large PagedLOD database with 0.3m imagery, and it stays at/above 60fps. Just another option for you to consider.
Ken. http://www.matrox.com/graphics/en/gxm/products/th2go/home.php On Mon, 2006-10-23 at 20:06 +0100, Robert Osfield wrote: > Hi Yefei, > > The best way to do three outputs from two cards is to have two render > surfaces, one per car, then have two camera/viewports sharing one > render surface on one card, and one camera/viewport on the other card. > > Robert. > > On 10/23/06, Yefei He <[EMAIL PROTECTED]> wrote: > > I use two GeForce 7950GX2 cards to drive three displays, running on > > Windows XP. Each card has two GPUs. It is able to render at 60Hz in > > a lot of situations. So it seems to be a single GPU, multi- > > rendersurface issue, not a single card, multi-rendersurface issue. > > I do notice some performance related issues though. > > > > 1. Rendering three channels is slower than rendering one or two > > channels (obviously). For one ive file I tested, I only got 30Hz > > when rendering three channels, but got 60Hz by rendering two > > channels, either one channel per card or even two channels driven by > > the same card. The antialiazing was set at 4x. When I dropped it to > > 2xQ, it was able to reach 60Hz with three channels. Limits in the > > throughput of the rendering pipeline? > > > > 2. What is the deal with cessna.osg? It takes over 10ms in the draw > > stage. dumptruck, cow, etc. all take less than half a ms. Dumptruck > > actually has a much higher triangle count. Needless to say, multi- > > channel configuration suffers when rendering cessna, as it takes > > over 10ms on each of the channels to draw after parts of the plane > > appear in all three channels. > > > > Yefei > > > > > > > > Date: Tue, 17 Oct 2006 09:00:49 -0700 > > > From: "Don Burns" <[EMAIL PROTECTED]> > > > Subject: Re: [osg-users] Multi-window performance > > > To: "osg users" <[email protected]> > > > Message-ID: > > > <[EMAIL PROTECTED]> > > > Content-Type: text/plain; charset="utf-8" > > > > > > The issue is two windows on one card, single threaded. This > > > should _not_ cause the performance to drop to 30 hz., but it > > > does... and probably because two swapBuffers are being issued > > > in the same frame. Since the driver is causing swapbuffers > > > to block, one swap has to wait for the next, causing the > > > frame rate to split in half. > > > > > > This also limits what we can do on the CPU to manage > > > threads/graphics contexts and frame rate, not to mention > > > keeping of statistics. > > > > > > I'd really like to get NVIdia to drop the blocking of the CPU > > > in the swapbuffers call, at least as an option for > > > applications that want to have better control in the CPU. > > > > > > -don > > > > > > > _______________________________________________ > > osg-users mailing list > > [email protected] > > http://openscenegraph.net/mailman/listinfo/osg-users > > http://www.openscenegraph.org/ > > > _______________________________________________ > osg-users mailing list > [email protected] > http://openscenegraph.net/mailman/listinfo/osg-users > http://www.openscenegraph.org/ _______________________________________________ osg-users mailing list [email protected] http://openscenegraph.net/mailman/listinfo/osg-users http://www.openscenegraph.org/
