Advanced search

Message boards : Graphics cards (GPUs) : since Power Cut always "Error while computing"

Author Message
ms113
Send message
Joined: 6 Feb 09
Posts: 19
Credit: 1,281,738
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 11805 - Posted: 11 Aug 2009 | 13:43:30 UTC

Dear all!

I am working on GPU-Grid since month w/o any remarkable issues. But ...

End of Juli my machine crashed due to a Power Cut.
And since then, after an ordinary reboot, I always get "Error while computing" processing GPU-Grid Tasks.
Details: Tasks

CPU-tasks (Einstein, Milkyway, a.s.o.) are running well.

I tried resetting the GPU-project, detaching it and reattaching it again, rebooting the machine and restarting the client-process ... nothing helps ...

.. further more, during testing I "always" run into the "daily quota gap" and so receive message(s) like e.g.:

Tue 11 Aug 2009 03:11:12 PM CEST|GPUGRID|Message from server: No work sent
Tue 11 Aug 2009 03:11:12 PM CEST|GPUGRID|Message from server: (reached daily quota of 3 tasks)

--> so testing is not really easy - concerning the timeframe.

Is there any Idea or help to solve this issue?

Thanks a lot in advance for any help and best regards,
Martin

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1957
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 11806 - Posted: 11 Aug 2009 | 14:04:29 UTC - in response to Message 11805.

what graphics driver do you have installed?

You need at least 185.xx.

gdf

ms113
Send message
Joined: 6 Feb 09
Posts: 19
Credit: 1,281,738
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 11807 - Posted: 11 Aug 2009 | 15:05:31 UTC - in response to Message 11806.
Last modified: 11 Aug 2009 | 15:06:26 UTC

I am currently using 177.82 on ubuntu 8.10

.. the latest nvidia-driver (available via package manager) is 180.11

???

should I use this one?
Do I have to update to Ubuntu 9.04 ?? (are there newer nvidia-drivers?)
Is there any other solution?
Why was it working until the "Crash by Power-Cut" end of July using Nvidia 177.82 driver?

Thanks for your help,
Martin

Profile GDF
Volunteer moderator
Project administrator
Project developer
Project tester
Volunteer developer
Volunteer tester
Project scientist
Send message
Joined: 14 Mar 07
Posts: 1957
Credit: 629,356
RAC: 0
Level
Gly
Scientific publications
watwatwatwatwat
Message 11813 - Posted: 11 Aug 2009 | 22:46:52 UTC - in response to Message 11807.

it is not working because we updated the Linux application to cuda2.1.
You should at least use 180.xx, better some newer driver (but it is a bit more difficult to upgrade).

gdf

RalphEllis
Send message
Joined: 11 Dec 08
Posts: 43
Credit: 2,216,617
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 11818 - Posted: 12 Aug 2009 | 6:00:41 UTC - in response to Message 11807.

You can manually upgrade your Nvidia driver. I have had some difficulty doing this with earlier versions of Ubuntu but here is a basic guide.
Go to
http://www.nvidia.com/object/cuda_get.html
It will guide you through the choices for Ubuntu 32 bit or 64 bit. Download the 190.18 beta driver. You can download the Cuda Toolkit but this is not absolutely necessary since cuda is in the driver.
Change the permissions on the downloaded file to make it executable.
In a terminal, go to the directory where the file is located.
type
sudo sh ./Nvidia-Linux*.pkg2.run -q --no-x-check

You will need to have the kernel sources installed and also gcc or the script will not work.

You may wish to save a backup copy of /etc/X11/xorg.conf
After the installation, type
sudo nvidia-xconfig
Afterwards, open the xorg.conf file to make sure that the driver section has
Driver "nvidia"
If you want to enable overclocking, put in the screen section
Option "Coolbits" "1"
To edit the xorg.conf file type
sudo gedit /etc/X11/xorg.conf

I am including a sample copy of my xorg.conf file
# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 1.0 (buildmeister@builder62) Wed Jul 22 16:45:17 PDT 2009

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"
HorizSync 28.0 - 33.0
VertRefresh 43.0 - 72.0
Option "DPMS"
EndSection

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

Section "Screen"
Option "Coolbits" "1"
Identifier "Screen0"
Device "Device0"
Monitor "Monitor0"
DefaultDepth 24
SubSection "Display"
Depth 24
EndSubSection
EndSection


Reboot the computer
The new driver should be enabled.
You will need to reset the Gpugrid project to get new projects.
If you upgrade your kernel at any point, you will need to run in a terminal
sudo sh ./NVIDIA-Linux*.pkg2.run -k --no-x-check
to update the Nvidia kernel module.

I have done this process many times in Suse. I have occasionally hit problems doing this in Ubuntu or Sabayon. I have done this process successfully in Ubuntu 9.04.
I have often found that if you restart the Gpugrid project after a driver install, the system will hang once. Don't be worried if this happens. Just reboot and reset the project.
Good luck
Ralph



ms113
Send message
Joined: 6 Feb 09
Posts: 19
Credit: 1,281,738
RAC: 0
Level
Ala
Scientific publications
watwatwatwatwat
Message 12237 - Posted: 31 Aug 2009 | 16:24:16 UTC

I've been on vacation - so could not respond yet.

Now finally, after updating to Ubuntu 9.04, using 180.44 it seems to work.

Thanks for your support.

Best regards,
Martin

Post to thread

Message boards : Graphics cards (GPUs) : since Power Cut always "Error while computing"

//