Setting up a machine with a Tesla K40 with Ubuntu 14.04 and an Intel on-board graphics card [solved]

Accelerated Computing CUDA CUDA Setup and Installation
1

I’m trying to install a system with the on-board graphics and a K40 card. After some tweaking I’ve gotten the on-board to work as the screen device but unfortunately there seems to be a problem with the nvidia-drivers. After installing CUDA 7.0 a conflict with the nvidia driver prevents unity from launching. Unfortunately the problem is not simply solved by running ccsm (compizconfig) and enabling the unity plugin as the OpenGL doesn’t work properly.

Here are the lspci relevant output:

  • 00:02.0 VGA compatible controller: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller (rev 06)
  • 01:00.0 3D controller: NVIDIA Corporation GK110BGL [Tesla K40c] (rev a1)

The issue started after installing cuda-repo-ubuntu1404-7-0-local_7.0-28_amd64.deb where the graphics defaulted to the K40 card even though the primary device is set to the on-board chip (according to the BIOS). I had to create a xorg.conf with the serverlayout specifying the intel device in order to get the graphics up and running again:

Section "ServerLayout"
    Identifier "layout"
    Screen 0 "intel"
    Inactive "nvidia"
EndSection

Section "Device"
    Identifier "intel"
    Driver "intel"
    BusID "PCI:0@0:2:0"
    Option "AccelMethod" "SNA"
EndSection

Section "Screen"
    Identifier "intel"
    Device "intel"
EndSection

Section "Device"
    Identifier "nvidia"
    Driver "nvidia"
    BusID "PCI:1@0:0:0"
    Option "ConstrainCursor" "off"
EndSection

Section "Screen"
    Identifier "nvidia"
    Device "nvidia"
    Option "AllowEmptyInitialConfiguration" "on"
    Option "IgnoreDisplayDevices" "CRT"
EndSection

In order to get the unity up and running I had to do a

sudo apt-get remove --purge nvidia*

After all this tweaking I’m where I started, with a basic Ubuntu setup… Is this conflict a known problem and if so, is there a work-around?

2

Ok, after another go at installing the the .deb file, removing driver blacklisting through the /etc/modprobes.d/nvidia-graphics-drivers.conf, using the aptitude setup it actually booted up nicely and the example files compile :-)

I also followed the instructions for the 6.5 install found here [url]http://www.r-tutor.com/gpu-computing/cuda-installation/cuda6.5-ubuntu[/url]

The aptitude solution solved a mesa-conflict, the solution is described here: [url]nvidia - Package dependency problem while installing CUDA on Ubuntu 14.04 - Ask Ubuntu

I hope this helps if you’re experiencing the same problem.

3

Are you able run any of the GUI examples?

(bindlessTexture Mandelbrot marchingCubes simpleGL simpleGLES simpleGLES_screen simpleTexture3D volumeFiltering volumeRender, nbody)

thanks in advance

4

I’m running into the same issue where I have an Alienware 13 with an onboard Intel graphics card and a NVIDIA GTX 960M and am running into the same issue where if I install cuda through the .deb file and then reboot, it boots to a black screen. In order to get rid of the black screen I must do:

sudo apt-get remove --purge nvidia*

@max_ki could you provide more details on what you mean by “removing driver blacklisting through the /etc/modprobes.d/nvidia-graphics-driver.conf”? I have installed CUDA 7.5 through

sudo aptitude install cuda

per your suggestion.

Any help would be appreciated.