FYI, I plan on doing this as well when I get my P1 in a few weeks. > I've installed the nvidia driver and disabled/blacklisted nouveau, however I > can't use bumblebee due it complaining there's no Intel GPU (Using `00:03.0 > VGA compatible controller: Device 1234:1111 (rev 02)`) >
My research led me to believe that you won't be able to use bumblebee with Qubes, as bumblebee expects both the Intel iGPU and Nvidia dGPU drivers in the same domain and Hybrid Graphics enabled in the BIOS. Since Qubes separates this by having the iGPU in dom0 (soon to be a new GUI domU?), and you are using pci-passthru dGPU to another domU, it won't have both. I plan on experimenting with this as well; however, I already know that the dGPU will always have to be active in this setup - killing any battery savings. What I don't know is if xen will put the dGPU will go into a power-savings mode if the domU it is attached to is shutdown. Or, better yet, if bumblebee can be installed in dom0 even with dGPU set as pci-passthru. I have my doubts; but, that is my plan to experiment with. My fallback is to disable the dGPU in the BIOS when on the road and switch back to Hybrid when only docked and only when I want to loadup AutoCAD or Blender (very rarely). However, I already know the X1 Exteme, and I'm guessing the P1, won't allow the Nvidia GPU to be disabled in the bios like earlier generation Thinkpads. Bleh, lots of workarounds. > So it's there, I just can't tell applications to render on the dGPU, so > close... > How can you tell "it's there"? Maybe look for it in the log such as: # grep "X Driver" /var/log/Xorg.0.log If you see something listed as Nvidia, then it is indeed loaded and could be used to offload hardware acceleration (such as Blender or perhaps AutoCAD shading). What I don't know is if rendering 3D to the screen would work within Qubes. OT: On my new Qubes/gaming desktop I am upgrading, I'll have all 4 displays connected to a single low-end GPU via HDMI inputs (monitors are currently all 1080p@120Hz). However, the 2x Vega 64s will connect via DVI (DP-to-DVI) to my primary 3 monitors for EyeFinity. A trick I saw someone do with Xen on youtube was to "disconnect" the monitors from HDMI of the first GPU - this frees up their signal. Then, within another domU, enable the DP output and BAM, the monitors come back - dedicated to the gaming domU. That should work with Qubes I believe with an HVM. However, I don't think that works with laptops though - unless you want to try dedicating the output of a specific external port to an external display. Maybe that might work? - keeping it disconnected from dom0, and enabling the output via xrandr within your nvidia dom0? -- You received this message because you are subscribed to the Google Groups "qubes-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to qubes-users+unsubscr...@googlegroups.com. To post to this group, send email to qubes-users@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/qubes-users/92324f5e-f474-45ef-93ee-bc5b41eaa6e4%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
