tapochek2004 Posted April 20, 2019 Share Posted April 20, 2019 Hello. I have a Dell Inspiron 7567 laptop with 2 GPUs: one is integrated Intel HD Graphics 630, and the second is discrete: Nvidia GeForce GTX 1050. The iGPU works normally with 1536 MB of VRAM, but I need HDMI working on this laptop. HDMI is wired directly to dGPU, which shows it is not Optimus. I was able to setup the dGPU to work on its own with external monitor connected in Windows and ArchLinux(via editing Xorg files). Here is the list of things I have already done: Updated to 10.12.6 Unchecked InjectNvidia in Clover Unchecked InjectIntel in Clover Installed Nvidia Web Drivers for Pascal Checked NvidiaWeb in Clover Installed Lilu.kext and NvidiaGraphicsFixup.kext Installed EmuVariable64 These steps made the dGPU show up in System Report as "Nvidia Chip Model" with 0MB of VRAM. I know it is possible to set up dGPU on this configuration, so I would like to get help. Link to comment Share on other sites More sharing options...
tapochek2004 Posted April 20, 2019 Author Share Posted April 20, 2019 If I check InjectNvidia in Clover, Nvidia shows up in System Report as Unknown, with Nvidia as vendor and 4095 MB of VRAM. If I try to enable nvda_drv=1 and InjectNvidia in Clover together, I get a kernel panic. Link to comment Share on other sites More sharing options...
tapochek2004 Posted April 20, 2019 Author Share Posted April 20, 2019 (edited) 5 hours ago, Hervé said: GeForce GTX1050 is Pascal and therefore requires the NVIDIA Web Driver since there is no native support for it (Apple do not provide kexts for any NVIDIA cards beyond Kepler generation). As such, you need the boot option nvda_drv=1 for the Web Driver to load. Don't think that, because HDMI is wired to the dGPU, you don't have Optimus technology. It's the exact opposite, you do, that's nVidia's solution for dual-GPU! I'm pretty certain that you need to enable Optimus in BIOS to use the iGPU and disable it to run only on the dGPU. That certainly is how it worked on many other Dell laptops. You can read about it at OSXLatitude... If you also use Lilu + WEG, you should not need to select "Inject NVIDIA" in your Clover config. Thanks, I understand. I am unable to disable iGPU in BIOS because it has no such function. I actually found a guy who was able to make 2 GPUs working on his Alienware laptop(https://github.com/acidanthera/WhateverGreen/blob/master/Manual/FAQ.GeForce.en.md) His guide suggests to make iGPU work in connector-less mode and use WhateverGreen (well that's how I understand his guide). I tried to follow it, but the problem is that WhateverGreen gives me KP unless I disable dGPU using SSDT patching. Could you please explain me if that guide could help in my situation and if I could do something to make WEG work? Edited April 20, 2019 by tapochek2004 Link to comment Share on other sites More sharing options...
Recommended Posts