Re: Two Nvidia cards, and black screen

[Arne G. G. <ar...@li...>]

* han...@sa...
> Hi,
>
> Well. This is now getting frustrating ;) I returned both NVIDIA cards 
> yesterday, after a long futile fight with them, but of course I never 
> tried the -sharevts option. Can you confirm that you have both NVIDIA 
> cards up and running with all features (TV-out, and 3D-accelleration)?

  $ DISPLAY=:0 glxinfo | grep OpenGL
  OpenGL vendor string: NVIDIA Corporation
  OpenGL renderer string: GeForce FX 5200/AGP/SSE2
  OpenGL version string: 2.0.0 NVIDIA 76.67
  OpenGL extensions:
  $ DISPLAY=:1 glxinfo | grep OpenGL
  OpenGL vendor string: NVIDIA Corporation
  OpenGL renderer string: GeForce FX 5200/PCI/SSE2
  OpenGL version string: 2.0.0 NVIDIA 76.67
  OpenGL extensions:
  $ glxgears 
  3773 frames in 5.0 seconds = 754.600 FPS
  3798 frames in 5.0 seconds = 759.600 FPS

I don't personally use the acceleration very heavily, so I've no idea
if the glxgears number is impressive or not.  At least the
acceleration seems to be enabled.  (The PCI card actually shows better
numbers than the AGP card.  Odd.)

I use S-video out only on the PCI card, but that works.  I believe I
have tested the AGP card and found that to work as well, without being
100% sure.

My only beef with the cards has been that I'm somehow unable to get
DVI out to work properly on the PCI card, but I have no reason to
suspect that this is caused by the multihead configuration.  (Nvidia
claims to have fixed some 1600x1200 DVI issues in the latest drivers,
but I haven't yet had the time to see if this makes any difference.)

> And can you also confirm that without the -sharevts option, you have a 
> similar screen blanking problem then the one described by me? If yes, I 
> just might go get some new NVIDIA cards ;)

Starting the second head X without the -sharevts option causes the
primary head to blank/poweroff, image returns when second head is
terminated.

As Cédric Rivard says, the -sharevts-option is an extension found (at
least) in the Ubuntu and Debian versions of X.org 6.8.2.  (Without
this option, I believe you need the ruby kernel patches to get around
precisely the problem seen above.)


							Arne.