Nvidia 319.12 beta for Linux, FreeBSD, and Solaris drivers with initial Optimus support

Finally, something happened that many have long been waiting for - Nvidia set their sights on and began to implement support for their Optimus hybrid graphics technology in * nix systems. At the moment, only a render is implemented on the Nvidia card, followed by forwarding the image to the Intel card. Dynamic change of the video card depending on the load is not yet supported .

The main changes compared to 313.30:

• Added new options for the Monitor section in xorg.conf: “Ignore”, “Enable”, “Primary” and “Rotate”. For example, the “Rotate” option can be used to automatically rotate output for a monitor with the specified EDID hash:
  
  

          Section "Monitor"
              Identifier "DPY-EDID-ee6cecc0-fa46-0c33-94e0-274313f9e7eb"
              Option "Rotate" "left"
          EndSection

  
  
• Implemented support for application profiles. You can learn more about how profiles work in the Application Profiles chapter of README.
• Added panning support for RandR 1.3.
• Added initial support for RandR 1.4, which added objects with the ability to select the source of the image output. This can be used to redirect output to another device or to organize several GPUs to work together when one GPU is used for output and the second for scene rendering (for example, output via DisplayLink USB adapter, and rendering to NVIDIA GPUs)
• Added the nvidia-modprobe utility, installed with setuid root privileges and used by driver components to control the loading of the kernel module and create the character device necessary for interacting with it. By default, standard distribution tools for loading modules and creating devices are still used, and the utility is positioned as a fallback when they are unavailable.
• The nvidia-settings command line utility has been updated, now you can specify output devices by name, for example:
  

  nvidia-settings -q [DVI-I-0]/RefreshRate
  nvidia-settings -q [GPU-1.DVI_I-1]/RefreshRate
  

  
  
• Also, now you do not need to specify “X screen 0” as the target, in case the operation or request does not require this. Instead, all devices will be processed.
• A page for configuring VDPAU parameters has been added to the NVIDIA control panel.
• Added support for dynamic video management through the RandR options: --newmode, --rmmode and --delmode of the xrandr utility.

Details here. A
draft guide on how to use the new features.