MultiGPU - no Xscreen for last GPU

Hello all,

I was on the geforce forum with my issue but they advised me to come here so here am I.

I have a 10 GPU system running under Ubuntu 16.04. GPUs are GTX 1060 3G cards from EVGA and MSI. I have only one monitor and it’s connected to the on board GPU, not to any of the GPUs.

I have configured everything in xorg.conf (see below) and everything works well except the last GPU. If I look at Nvidia X server settings, I see all GPUs (from 0 to 9) but I only see 9 corresponding X screen. I have enabled coolbits 12 to all of the cards but for some reason, it seems to be completely ignored for the last GPU. The consequence is that I cannot control fan and memory speed for the last GPU (while it works with all the other ones).

My xorg.conf file looks as follows:
Section “ServerLayout”
Identifier “layout”
Screen 0 “screen1” 0 0
Screen 1 “screen2” 0 0
Screen 2 “screen3” 0 0
Screen 3 “screen4” 0 0
Screen 4 “screen5” 0 0
Screen 5 “screen6” 0 0
Screen 6 “screen7” 0 0
Screen 7 “screen8” 0 0
Screen 8 “screen9” 0 0
Screen 9 “screen10” 0 0
Inactive “intel”
InputDevice “Keyboard0” “CoreKeyboard”
InputDevice “Mouse0” “CorePointer”
EndSection

Section “InputDevice”
# generated from default
Identifier “Keyboard0”
Driver “keyboard”
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 “Monitor”
Identifier “Monitor0”
VendorName “Unknown”
ModelName “Unknown”
HorizSync 28.0 - 33.0
VertRefresh 43.0 - 72.0
Option “DPMS”
EndSection

Section “Device”

#Driver “intel”
Identifier “intel”
Driver “modesetting”
Option “AccelMethod” “None”
#Option “AccelMethod” “sna”
BusID “PCI:0@0:2:0”
EndSection

Section “Device”
Identifier “nvidia1”
Driver “nvidia”
BusID “PCI:1:0:0”
EndSection

Section “Device”
Identifier “nvidia2”
Driver “nvidia”
BusID “PCI:2:0:0”
EndSection

Section “Device”
Identifier “nvidia3”
Driver “nvidia”
BusID “PCI:3:0:00”
EndSection

Section “Device”
Identifier “nvidia4”
Driver “nvidia”
BusID “PCI:4:0:0”
EndSection

Section “Device”
Identifier “nvidia5”
Driver “nvidia”
BusID “PCI:5:0:0”
EndSection

Section “Device”
Identifier “nvidia6”
Driver “nvidia”
BusID “PCI:8:0:0”
EndSection

Section “Device”
Identifier “nvidia7”
Driver “nvidia”
BusID “PCI:10:0:0”
EndSection

Section “Device”
Identifier “nvidia8”
Driver “nvidia”
BusID “PCI:11:0:0”
EndSection

Section “Device”
Identifier “nvidia9”
Driver “nvidia”
BusID “PCI:12:0:0”

EndSection

Section “Device”
Identifier “nvidia10”
Driver “nvidia”
BusID “PCI:13:0:0”

EndSection

Section “Screen”
Identifier “intel”
Device “intel”
Monitor “Monitor0”
EndSection

Section “Screen”
Identifier “screen1”
Device “nvidia1”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen2”
Device “nvidia2”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen3”
Device “nvidia3”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen4”
Device “nvidia4”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen5”
Device “nvidia5”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen6”
Device “nvidia6”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen7”
Device “nvidia7”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen8”
Device “nvidia8”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen9”
Device “nvidia9”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

Section “Screen”
Identifier “screen10”
Device “nvidia10”
Monitor “Monitor0”
DefaultDepth 24
Option “AllowEmptyInitialConfiguration” “on”
Option “IgnoreDisplayDevices” “CRT”
Option “ConstrainCursor” “off”
Option “Coolbits” “12”
Option “MultiGPU” “on”
SubSection “Display”
Depth 24
Modes “nvidia-auto-select”
EndSubSection
EndSection

As you can see, all GPUs are configured exactly the same way but still they don’t behave the same way.

I’m using driver version 384.69, I also tried older version but that doesn’t change anything. The whole setup used to work as expected but after a recent update, I couldn’t change settings for the last GPUs anymore.

Do you have any idea of what may have caused this?

Thanks
Laurent

>>after a recent update, I couldn’t change settings for the last GPUs anymore.
What did you changed ? Please attach nvidia report . Make sure there is not any hardware issue with that gpu.

Hello Sandip,

By update I mean a system update, I did not pay much attention to the different component in the batch, I just wanted to have a updated version of Ubuntu.

You can find the bug report here: laurentwillen.be/nvidia-bug-report.log.gz

There is no hardware issue with the GPU, it’s running perfectly well, the only issue is that I can’t overclock it for some reason.

Thanks!
Laurent

Hello Sandip,

have you got any news?

Thanks

You have two xorg.conf, one in /etc/X11 and one in /etc/X11/xorg.conf.d, please delete the latter one and run nvidia-bug-report.sh again. It’s probably not the reason for your problem, but it’s hard to tell which settings are effectively used.

Hello Generic,

Brillant! I had not seen this and now that I have removed the file and rebooted, I now have access to all my cards!

Thanks for your help!!