2

(I am aware that questions about doing this specific task have already been answered numerous time here. However there doesn't seem to have much information for this specific version of Ubuntu, and having myself tried multiple methods unsuccessfully, I think creating an entry here is necessary.)

Description

I would like to use my desktop computer's NVIDIA GPU to debug and run some CUDA programs. Since software preemption is a little deceiving, I decided I'd switch display to my unused Intel iGPU and use the NVIDIA card exclusively for the CUDA computations.

I'd like to know what would be the most efficient way to do so with my current Ubuntu and CUDA version.

I'd also like if possible to switch the display the easiest way back to NVIDIA when I'm done.

Attempted methods

  • Switching profiles using NVIDIA PRIME from NVIDIA X server settings isn't possible as the entry isn't available
  • Using prime-select intel crashed Ubuntu on next boot, had to chroot into it and launch prime-select nvidia to fix*

* Attempted with a slightly older NVIDIA driver version.

Additional information

lspci | egrep 'VGA|3D'

00:02.0 VGA compatible controller: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller (rev 06)
01:00.0 VGA compatible controller: NVIDIA Corporation GM107 [GeForce GTX 750 Ti] (rev a2)

cat /etc/X11/xorg.conf

# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig:  version 455.45.01

Section "ServerLayout"
    Identifier     "Layout0"
    Screen      0  "Screen0"
    InputDevice    "Keyboard0" "CoreKeyboard"
    InputDevice    "Mouse0" "CorePointer"
EndSection


Section "Files"
EndSection


Section "InputDevice"
    # generated from default
    Identifier     "Mouse0"
    Driver         "mouse"
    Option         "Protocol" "auto"
    Option         "Device" "/dev/psaux"
    Option         "Emulate3Buttons" "no"
    Option         "ZAxisMapping" "4 5"
EndSection


Section "InputDevice"
    # generated from default
    Identifier     "Keyboard0"
    Driver         "kbd"
EndSection


Section "Monitor"
    Identifier     "Monitor0"
    VendorName     "Unknown"
    ModelName      "Unknown"
    Option         "DPMS"
EndSection


Section "Device"
    Identifier     "Device0"
    Driver         "nvidia"
    VendorName     "NVIDIA Corporation"
EndSection


Section "Screen"
    Identifier     "Screen0"
    Device         "Device0"
    Monitor        "Monitor0"
    DefaultDepth    24
    SubSection     "Display"
        Depth       24
    EndSubSection
EndSection
  • Ubuntu version: 20.04
  • Kernel Version: 5.4.0-57-generic
  • NVIDIA Driver version: 455.45.01
  • CUDA version: 11.1
  • NVIDIA GPU: GTX 750Ti
  • Intel iGPU: Intel 4600HD
CodeTalker
  • 21

1 Answers1

-1

nvidia-smi may be what you're looking for.

sudo apt install nvidia-smi

Power management mode settings

nvidia-smi -i <device id> -pl N
sudo nvidia-persistenced
nvidia-smi -q -d SUPPORTED_CLOCKS$ sudo nvidia-smi -ac <Mem clock, Graphics clock>

nvidia-smi commands: https://subscription.packtpub.com/book/programming/9781788996242/app01/app01sec01/useful-nvidia-smi-commands

nvidia-smi repositories: https://github.com/topics/nvidia-smi

https://manpages.ubuntu.com/manpages/precise/man1/alt-nvidia-current-smi.1.html

http://http.download.nvidia.com/developer/GPU_Programming_Guide/GPU_Programming_Guide.pdf

яєν.unkn0wn
  • 1
  • As far as I read, nvidia-smi commands are mainly about obtaining information about GPU and monitoring. I can't see how the few concrete actions you can have on using it solve the problem here – CodeTalker Dec 14 '20 at 18:49