Nvidia-Ubuntu 14.04 fail with latest kernel

Message boards : Number crunching : Nvidia-Ubuntu 14.04 fail with latest kernel
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1795412 - Posted: 11 Jun 2016, 16:53:03 UTC

Installed the latest (weekly lately) 64bit kernel on CPUid 5766757
Ubuntu 14.04. Did the sequence that has worked for several months
after getting a kernel update:

sudo apt-get purge 'nvidia*'
sudo shutdown -h now
reboot....
Use settings additional drivers to get nvidia
sudo shutdown -h now
reboot...
sudo apt-get install nvidia-modprobe


It's a very old nvidia-modprobe, but oh well.
X won't start. No graphics.

If I don't install nvidia-modprobe graphics work
but boinc won't find the cards... And yes, I do restart
boinc-client, but that does not help boinc find the cards.

So I aborted the Seti GPU tasks in hand. A lovely bunch of
GPU tasks that I have no way to process right now...
Doing cpu for now on 11 cores.
ID: 1795412 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1795440 - Posted: 11 Jun 2016, 18:29:39 UTC

I meant Xubuntu, not plain Ubuntu.
16.04 will be officially
marked LTS in the repositories in a month
or so and when it is I'll upgrade.
In hopes it will help with
GPU on boinc. Sensible to do anyway.
ID: 1795440 · Report as offensive
TBar
Volunteer tester

Send message
Joined: 22 May 99
Posts: 5204
Credit: 840,779,836
RAC: 2,768
United States
Message 1795443 - Posted: 11 Jun 2016, 18:46:54 UTC - in response to Message 1795440.  

It's probably just another problem with the repository driver. I'm using 364.19 from nVidia and it seems to be working fine with Ubuntu 3.13.0-88;
http://setiathome.berkeley.edu/show_host_detail.php?hostid=7258715
ID: 1795443 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1795462 - Posted: 11 Jun 2016, 21:43:58 UTC

I managed to install the official NVIDIA latest driver to my ubuntu 15 for the first time ever. I had to apt-get purge/remove all nvidia*, uninstall lighdm, try kde and uninstall that and then install lightdm and nvidia drivers. It took me a week and I did not write down everything I did. But it is doable.

Petri (Pee Tree)
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1795462 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1796480 - Posted: 16 Jun 2016, 1:57:28 UTC

My concern is that it is (apparently) the old nvidia-modprobe
that (in 14.04) kills X/lightdm. Why that has not been
updated in sync with the drivers (in ubuntu) is a puzzle.
Someone thinks it's not needed?
Do you have nvidia-modprobe petri33?

Without nvidia-modprobe
the desktop (xfce) comes up fine with either nvidia
driver that 'additional drivers' shows. But boinc cannot see
the GPUs.
ID: 1796480 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1799360 - Posted: 29 Jun 2016, 14:16:22 UTC

Updated kernel installed June 28 and installed and ran nvidia-modprobe (same one as before) and X/lightdm stay operational.
So GPU operations have been resumed.

a fair number of GPU tasks validated overnight.

On one task 5008182283 I got:
ERROR: Possible wrong computation state on GPU, host needs reboot or maintenance

which is alarming. I'll be watching all this closely.
ID: 1799360 · Report as offensive
The_Matrix
Volunteer tester

Send message
Joined: 17 Nov 03
Posts: 414
Credit: 5,827,850
RAC: 0
Germany
Message 1799441 - Posted: 29 Jun 2016, 19:43:25 UTC - in response to Message 1796480.  
Last modified: 29 Jun 2016, 20:16:17 UTC

Someone thinks it's not needed?


I prefer Ubuntu 15.10.

And this no gpu is found could be solved with that:

https://boinc.berkeley.edu/dev/forum_thread.php?id=6307

ok, here is a shortcut installion instruction from me :)

http://setiathome.berkeley.edu/forum_thread.php?id=79358
ID: 1799441 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1799493 - Posted: 29 Jun 2016, 22:20:28 UTC - in response to Message 1799441.  

Thanks for the links, The_Matrix.
ID: 1799493 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1799726 - Posted: 30 Jun 2016, 19:01:50 UTC

The failed task I mentioned yesterday is failed for 4 others (3 have GPU related messages suggesting reboot same as I got, 1 has complaint about too-old GPU.) So I guess that task failure is nothing I should worry about.
ID: 1799726 · Report as offensive
Profile Empire_Builder
Avatar

Send message
Joined: 4 Sep 11
Posts: 17
Credit: 753,579
RAC: 0
United States
Message 1799735 - Posted: 30 Jun 2016, 20:23:02 UTC

So far my experience with nVidia drivers on Linux is you get it to work once, then don't touch anything again.
ID: 1799735 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1799786 - Posted: 1 Jul 2016, 0:10:47 UTC - in response to Message 1799735.  

I prefer to keep up with bug fixes. Only one machine
has the nVidia driver problem, and that is only with some kernel releases
in 14.04 LTS. So I occasionally struggle :-)

Two nvidia GPU machines here on 14.04 are updated too
and just work without surprises.
ID: 1799786 · Report as offensive
Profile David Anderson (not *that* DA) Project Donor
Avatar

Send message
Joined: 5 Dec 09
Posts: 215
Credit: 74,008,558
RAC: 74
United States
Message 1802796 - Posted: 15 Jul 2016, 23:38:56 UTC

New Linux kernel today. Install went fine, GPUs fine,
Seti working fine.
ID: 1802796 · Report as offensive
Profile ML1
Volunteer moderator
Volunteer tester

Send message
Joined: 25 Nov 01
Posts: 20147
Credit: 7,508,002
RAC: 20
United Kingdom
Message 1803270 - Posted: 18 Jul 2016, 11:20:12 UTC - in response to Message 1802796.  
Last modified: 18 Jul 2016, 11:23:01 UTC

New Linux kernel today. Install went fine, GPUs fine,
Seti working fine.

Good that a new kernel worked.


For reference for others:

Which kernel version did you need to move up to to get it working again?

(I guess a 2014 distro is nearing end of support? Or is that one a "LTS"?)


Happy cool crunchin',
Martin


(A quick way to find out is to open a command line and give the command:

uname -a

to show the system name and kernel name.)
See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)
ID: 1803270 · Report as offensive

Message boards : Number crunching : Nvidia-Ubuntu 14.04 fail with latest kernel


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.