This might be slightly off what your final question was, but I think it is pertinent to the situation. If you are struggling with NVidia drivers, this link has been super helpful for me.
http://www.terriblesysadmin.com/?p=38
Basically, these are the steps you need to do:
sudo apt-get update
sudo apt-get remove --purge nvidia-* xserver* libcheese7
sudo apt-get install xserver-xorg-lts-xenial
sudo apt-get install nvidia-384 nvidia-settings nvidia-prime ubuntu-desktop
sudo reboot
Notice that the driver is nvidia-384
instead of the older driver.
Now, if you finish to this point, you should have a stable machine.
BUT
I highly recommend, if you plan on doing CUDA or other GPU heavy swag, that you install CUDA 10.1 from dpkg
instead. Installing CUDA not only installed CUDA (shocker), but it also installs the video driver that is most up-to-date and compatible with your system, which I believe is 418
. Basically, you need to boot to a recovery (root) prompt and then run the associated sudo dpkg -i theinstaller.deb
.
These are the local deb instructions:
sudo dpkg -i cuda-repo-ubuntu1604-10-1-local-10.1.168-418.67_1.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-<version>/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda
From this link https://developer.nvidia.com/cuda-downloads
. I recommend the local deb
link because it has been the most stable, consistent solution. I can also navigate the the NVidia files this way: the run
file does not work for me.
I should also mention: the Acer Nitro 5, no matter what model (53, 55, you name it), will always kick the can when you install linux for the first time. You are not the only one struggling with this process. After installing the CUDA provided driver, I have not had any issues since with my Nitro 5.
One last thing:
About the gdm3
and lightdm
: those are GUI managers. They are service
s that you can start and stop. compiz
is also a common GUI manager. I am not an expert in GUI managers, so I would recommend you search those 3 specific terms. For the Nvidia driver installation of this answer, I am %100 confident.
On a side note: it appears that installing the NVidia drivers also gets the cooling fans to work properly. When I am running an intense Keras neural network training script, the GPU cooling fans kick in. More reason to simply install CUDA...