Hello,
I recently reconfigured my main desktop to a dual GTX460 configuration and came across several references that say that I can not use SLI and two monitors at the same time. This combination has been supported on windows for at least 4 years. Is there any intention of giving the linux driver feature parity in this case?
I had on linux SLI + 2 monitors, but both were connected at the same card. Check this page as well: NVIDIA GeForce Graphics Cards: The Ultimate in PC Gaming
pasoleatis: Looks like I must bang my head harder against my xorg.conf as nothing relevant is in the 313.30 nvidia-settings or random forums. Is there any way to confirm that page is accurate for Fermi/Kepler GPUs?
I can confirm I had 2 monitors. I had a 2 x 660 Ti in SLi configuration. I did not do anything special, it just work. I put one on the dvi (with a dvi–>vga converter) and one on the hdmi. No special setups. Ubuntu 12.04 + 313 driver and/or 319.
Thanks for the confirmation.
Turns out I needed to put Option “Coolbits” “18” in order to override a vram size mismatch. That gives me the lesser video card claiming it has the same vram as the better videocard (bad behavior) and this:
[ 21.233] (WW) NVIDIA(0): Multiple display devices requested in MetaMode
[ 21.233] (WW) NVIDIA(0): “DFP-0:nvidia-auto-select,DFP-2:nvidia-auto-select”, but
[ 21.233] (WW) NVIDIA(0): only one allowed when SLI is enabled; ignoring extra
[ 21.233] (WW) NVIDIA(0): display devices.
Also, gdm-3.6.2 falls over itself and gives me a cursor over a black screen.
I don’t know if it was changed in Kepler, and I can’t say about Fermi, but I remember on the 7950 GX2 I could not do both SLI and TwinView. I had to choose one or the other. It was documented fairly commonly on the old nvnews forum. I think it was like a Quadro-only feature or something like that.