Goal

I am trying to use CUDA on my nvidia card for research. I don't really care to use it to manage my display as I plan on only using the computer via bash-shell after I finish setting it up.

Problem

My video card is unclaimed by Ubuntu. Bounce to login loop after signing in.

Background

I'm a linux-savy, power-user, computer science phd student, but I'm stumped trying to get my Nvidia gtx 1070Ti graphics card to work. I've been at this every sunday for over two months now.

I've followed these tutorials:

https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia
https://help.ubuntu.com/community/BinaryDriverHowto
https://kislayabhi.github.io/Installing_CUDA_with_Ubuntu/
https://askubuntu.com/a/760935/13693
https://askubuntu.com/a/937204/13693
http://docs.nvidia.com/cuda/cuda-installation-guide-linux

Installing nvidia-current or nvidia-387 (default chosen when ubuntu installed) , or the latest nvidia-390 results in a boot loop where I'm bounced back to the login screen after login in.

So I used prime-select intel and removed the modeset=0 blacklist to get to a working desktop. So below is a review of my current status:

$ uname -a
Linux datalake2 4.13.0-36-generic #40~16.04.1-Ubuntu SMP Fri Feb 16 23:25:58 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
$ lspci | grep VGA
03:00.0 VGA compatible controller: NVIDIA Corporation Device 1b82 (rev a1)
08:00.0 VGA compatible controller: Matrox Electronics Systems Ltd. G200eR2 (rev 01)
$ sudo lshw -C video
 *-display UNCLAIMED
 description: VGA compatible controller
 product: NVIDIA Corporation
 vendor: NVIDIA Corporation
 physical id: 0
 bus info: pci@0000:03:00.0
 version: a1
 width: 64 bits
 clock: 33MHz
 capabilities: pm msi pciexpress vga_controller cap_list
 configuration: latency=0
 resources: iomemory:33f0-33ef iomemory:33f0-33ef memory:91000000-91ffffff memory:33fe0000000-33fefffffff memory:33ff0000000-33ff1ffffff ioport:2000(size=128) memory:92080000-920fffff

$ apt list --installed | grep "nvidia"

nvidia-387/unknown,now 387.26-0ubuntu1 amd64 [installed]
nvidia-387-dev/unknown,now 387.26-0ubuntu1 amd64 [installed,automatic]
nvidia-cuda-dev/xenial,now 7.5.18-0ubuntu1 amd64 [installed,automatic]
nvidia-cuda-doc/xenial,xenial,now 7.5.18-0ubuntu1 all [installed,automatic]
nvidia-cuda-gdb/xenial,now 7.5.18-0ubuntu1 amd64 [installed,automatic]
nvidia-cuda-toolkit/xenial,now 7.5.18-0ubuntu1 amd64 [installed]
nvidia-modprobe/unknown,now 387.26-0ubuntu1 amd64 [installed,automatic]
nvidia-opencl-dev/xenial,now 7.5.18-0ubuntu1 amd64 [installed,automatic]
nvidia-opencl-icd-387/unknown,now 387.26-0ubuntu1 amd64 [installed,automatic]
nvidia-prime/xenial,now 0.8.2 amd64 [installed]
nvidia-profiler/xenial,now 7.5.18-0ubuntu1 amd64 [installed,automatic]
nvidia-settings/unknown,now 387.26-0ubuntu1 amd64 [installed,automatic]
nvidia-visual-profiler/xenial,now 7.5.18-0ubuntu1 amd64 [installed,automatic]

$ cat /proc/driver/nvidia/version
cat: /proc/driver/nvidia/version: No such file or directory

Weirdness

My second problem seems to be that ubuntu is unable to recognize the need for drivers for my card, even-though I have enabled restricted propitiatory drivers.

sudo software-properties-gtk gives me nothing as well.

My gcc version:

Here is the workaround:

1. edit /etc/default/grub

Modify GRUB_CMDLINE_LINUX_DEFAULT to

GRUB_CMDLINE_LINUX_DEFAULT='pcie_port_pm=off acpi_backlight=none acpi_osi=Linux acpi_osi=! acpi_osi="Windows 2009"'

This step is to prevent blank screen after logging in.

2. move nvidia library directories to /etc/ld.so.conf.d/nvidia.conf

The content of nvidia.conf is

/usr/lib/nvidia-390
/usr/lib32/nvidia-390

These directories depends on driver version on your computer.

3. create /etc/init.d/nvidia

To disable and enable nvidia runtime libraries.

#!/bin/sh
### BEGIN INIT INFO
# Provides: nvidia 
# Required-Start: $all
# Required-Stop: $all
# Default-Start: 5
# Default-Stop: 0 6
# Short-Description: load/unload nvidia library
# Description: load/unload nvidia library
### END INIT INFO

PRIME=$(prime-select query)
if [ "$PRIME" = "nvidia" ]; then
 exit 0
fi

case "$1" in
 start)
 sleep 10
 cd /etc/ld.so.conf.d
 mv nvidia.conf.bak nvidia.conf
 ldconfig
 nvidia-smi
 ;;
 stop)
 cd /etc/ld.so.conf.d
 mv nvidia.conf nvidia.conf.bak
 ldconfig
esac

4. execute update-rc.d nvidia defaults

You should find SXXnvidia in /etc/rc5.d/ and KXXnvidia in /etc/rc6.d/, /etc/rc0.d/.

Try to execute /etc/init.d/nvidia stop and nvidia-smi, you should see error messages of libraries not found.

Try to execute /etc/init.d/nvidia start, then nvidia-smi is fine again.

If everything is OK, you can reboot now. You are expected to login to desktop.

5. If anything goes wrong

The most possible problem is nvidia script not executed. If it happens, you can press Ctrl+Alt+F1 to tty mode, execute /etc/init.d/nvidia stop; reboot. Then you can go back to unity desktop to debug.

6. known side-effect

When use intel as prime GPU, unity-control-center(system settings) will be failed to start.

GLib-CRITICAL **: g_strsplit: assertion `string != NULL' failed.

Note: my system spec

# uname -r
4.13.0-32-generic
# lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 16.04.3 LTS
Release: 16.04
Codename: xenial
# dpkg -l | grep cuda
ii cuda-9-0 9.0.176-1 amd64 CUDA 9.0 meta-package
ii cuda-command-line-tools-9-0 9.0.176-1 amd64 CUDA command-line tools
ii cuda-core-9-0 9.0.176-1 amd64 CUDA core tools
ii cuda-cublas-9-0 9.0.176.1-1 amd64 CUBLAS native runtime libraries
ii cuda-cublas-dev-9-0 9.0.176.1-1 amd64 CUBLAS native dev links, headers
ii cuda-cudart-9-0 9.0.176-1 amd64 CUDA Runtime native Libraries
ii cuda-cudart-dev-9-0 9.0.176-1 amd64 CUDA Runtime native dev links, headers
ii cuda-cufft-9-0 9.0.176-1 amd64 CUFFT native runtime libraries
ii cuda-cufft-dev-9-0 9.0.176-1 amd64 CUFFT native dev links, headers
ii cuda-curand-9-0 9.0.176-1 amd64 CURAND native runtime libraries
ii cuda-curand-dev-9-0 9.0.176-1 amd64 CURAND native dev links, headers
ii cuda-cusolver-9-0 9.0.176-1 amd64 CUDA solver native runtime libraries
ii cuda-cusolver-dev-9-0 9.0.176-1 amd64 CUDA solver native dev links, headers
ii cuda-cusparse-9-0 9.0.176-1 amd64 CUSPARSE native runtime libraries
ii cuda-cusparse-dev-9-0 9.0.176-1 amd64 CUSPARSE native dev links, headers
ii cuda-demo-suite-9-0 9.0.176-1 amd64 Demo suite for CUDA
ii cuda-documentation-9-0 9.0.176-1 amd64 CUDA documentation
ii cuda-driver-dev-9-0 9.0.176-1 amd64 CUDA Driver native dev stub library
ii cuda-drivers 390.12-1 amd64 CUDA Driver meta-package
ii cuda-libraries-9-0 9.0.176-1 amd64 CUDA Libraries 9.0 meta-package
ii cuda-libraries-dev-9-0 9.0.176-1 amd64 CUDA Libraries 9.0 development meta-package
ii cuda-license-9-0 9.0.176-1 amd64 CUDA licenses
ii cuda-misc-headers-9-0 9.0.176-1 amd64 CUDA miscellaneous headers
ii cuda-npp-9-0 9.0.176-1 amd64 NPP native runtime libraries
ii cuda-npp-dev-9-0 9.0.176-1 amd64 NPP native dev links, headers
ii cuda-nvgraph-9-0 9.0.176-1 amd64 NVGRAPH native runtime libraries
ii cuda-nvgraph-dev-9-0 9.0.176-1 amd64 NVGRAPH native dev links, headers
ii cuda-nvml-dev-9-0 9.0.176-1 amd64 NVML native dev links, headers
ii cuda-nvrtc-9-0 9.0.176-1 amd64 NVRTC native runtime libraries
ii cuda-nvrtc-dev-9-0 9.0.176-1 amd64 NVRTC native dev links, headers
ii cuda-repo-ubuntu1604 9.1.85-1 amd64 cuda repository configuration files
ii cuda-runtime-9-0 9.0.176-1 amd64 CUDA Runtime 9.0 meta-package
ii cuda-samples-9-0 9.0.176-1 amd64 CUDA example applications
ii cuda-toolkit-9-0 9.0.176-1 amd64 CUDA Toolkit 9.0 meta-package
ii cuda-visual-tools-9-0 9.0.176-1 amd64 CUDA visual tools
ii libcuda1-390 390.12-0ubuntu1 amd64 NVIDIA CUDA runtime library
ii libcudnn7 7.0.5.15-1+cuda9.0 amd64 cuDNN runtime libraries
ii libcudnn7-dev 7.0.5.15-1+cuda9.0 amd64 cuDNN development libraries and headers
# dpkg -l | grep nvidia
ii nvidia-390 390.12-0ubuntu1 amd64 NVIDIA binary driver - version 390.12
ii nvidia-390-dev 390.12-0ubuntu1 amd64 NVIDIA binary Xorg driver development files
ii nvidia-modprobe 390.12-0ubuntu1 amd64 Load the NVIDIA kernel driver and create device files
ii nvidia-opencl-icd-390 390.12-0ubuntu1 amd64 NVIDIA OpenCL ICD
ii nvidia-prime 0.8.2 amd64 Tools to enable NVIDIA's Prime
ii nvidia-settings 390.12-0ubuntu1 amd64 Tool for configuring the NVIDIA graphics driver

You should be able to get CUDA working with this answer. by Ping Chu Hung If you still have issues with the login loop after that there are some highly rated answers here that should resolve that for you.

Note: Like most things in life, Nvidia drivers can leave a bunch of garbage lying around if you've tried to install several versions or had failed installations and it may be necessary to purge them all and then reinstall the one you've had working in the past to get the desired results.