LinuxQuestions.org
Visit the LQ Articles and Editorials section
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Hardware
User Name
Password
Linux - Hardware This forum is for Hardware issues.
Having trouble installing a piece of hardware? Want to know if that peripheral is compatible with Linux?

Notices

Reply
 
Search this Thread
Old 03-07-2012, 04:44 PM   #16
odiseo77
Member
 
Registered: Dec 2004
Location: London, UK
Distribution: Debian Sid, OpenSUSE 13.1
Posts: 995

Original Poster
Rep: Reputation: 298Reputation: 298Reputation: 298

Maybe my problem lies somewhere else because the type of freezes are different, although I could try to switch to nouveau or an earlier version of the nvidia proprietary driver to see if the problem stops. Thanks for the heads up.

BTW, I still haven't bought the DVM (haven't got the time yet). Will buy it in the upcoming days and let you folks know of any news I find.

Thanks all for the help!

Last edited by odiseo77; 03-07-2012 at 04:50 PM.
 
Old 03-07-2012, 04:57 PM   #17
afreitascs
Member
 
Registered: Aug 2004
Location: Brasil
Distribution: Slackware_Cur-64_mult
Posts: 433

Rep: Reputation: 30
Classic hardware problem freezing:

CPU temperature ...

Compatibility of memory (not your case)

PSU (use another psu to test)

video card (so try using the onboard video)

you have to separate the problem if the problem is hardware or software.

good luck ( buena suerte ) boa sorte !
 
Old 03-08-2012, 05:23 AM   #18
cascade9
Senior Member
 
Registered: Mar 2011
Location: Brisneyland
Distribution: Debian, aptosid
Posts: 3,718

Rep: Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903
Quote:
Originally Posted by odiseo77 View Post
I checked the heatsink and though it's a little bit dusty, I don't think it's as dusty as to cause any malfunction. I also used cpuburn (a cpu stress program) for about 5 or 8 minutes, checked the CPU temperature and they didn't rise too much during this time; from 38 C idle to around 52 C under load, although it probably would have rised more if I had let cpuburn run longer; (the machine didn't freeze during the stress test).
Thats pretty hot to get in a few minutes, but not hot enough that I think you've problems are caused by overheating. One down, tons more to go...

Quote:
Originally Posted by odiseo77 View Post
Now I'm starting to be suspcicious about the PSU. I replaced the one I had before because the machine just stopped working one day (it didn't even beeped, no leds, the PSU fan didn't spin, etc.); so I took this one I had bought like 2 years ago but had not used, connected it, and the PC worked again. Will have to do the voltage test these days to see what I find. BTW; in case I decide -or have- to buy a new PSU, which brands would you recommend?
I'm a sucker for seasonic. Not cheap, but I've never had one fail. Corsair is pretty good as well, mostly they use seasonic PSUs (rebranded to corsair), but the ones that aren't seasonic based are good as well.

Quote:
Originally Posted by odiseo77 View Post
Maybe my problem lies somewhere else because the type of freezes are different, although I could try to switch to nouveau or an earlier version of the nvidia proprietary driver to see if the problem stops. Thanks for the heads up.
Switch to nouveau. If that doesnt fix it, remove the nVidia card and try the onboard video.

If the onboard works, that doesnt mean that nvidia-glx and nouveau dont work, it could be a sick/dying video card. Better than than motherboard issues.

BTW, heres a few other things you can try-

Go into your BIOS and drop the memroy speeds, disable all the stuff you arent using, eg the floppy drive, serial/parallel port, IDE/SATA ports (if not is use) etc.. BTW, even if disabling useless/uneeded stuff in the BIOS doesnt seem to help, IMO its always worth doing. The more useless/uneeded stuff enabled in BIOS, the more time it takes to load, more shared IRQs, and so on.

If that fails, try removing 1 stick of RAM, all the HDDs/optical drives excpet for the boot drive, and disable even more in the BIOS, even stuff you use/need (eg sound, network) and try again. I dont think that you have a RAM problem, but once you've tried the intel video, slowed the RAM down and disabled virtually everything in the BIOS, you're running out of easy things to test. So running at the most 'cut down' you can is worth trying.
 
1 members found this post helpful.
Old 03-08-2012, 06:01 AM   #19
sycamorex
LQ Veteran
 
Registered: Nov 2005
Location: London
Distribution: Slackware64-current
Posts: 5,564
Blog Entries: 1

Rep: Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026
Quote:
Originally Posted by cascade9 View Post
If that fails, try removing 1 stick of RAM, all the HDDs/optical drives excpet for the boot drive, and disable even more in the BIOS, even stuff you use/need (eg sound, network) and try again. I dont think that you have a RAM problem, but once you've tried the intel video, slowed the RAM down and disabled virtually everything in the BIOS, you're running out of easy things to test. So running at the most 'cut down' you can is worth trying.
That reminds me of a problem I had last year. I bought some fancy RAM (over 2000MHz). I could only set the BIOS to 1866MHz to be able to boot the machine without it freezing straight away. Fair enough. Still I noticed that while doing some CPU intensive tasks (eg. compiling a kernel with the -j5/7 flag the system would freeze. What turned out to be the problem was that my CPU only supports RAM up to 1066MHz. I hadn't known that there's any correlation between a cpu and RAM frequency.
 
Old 03-08-2012, 03:48 PM   #20
odiseo77
Member
 
Registered: Dec 2004
Location: London, UK
Distribution: Debian Sid, OpenSUSE 13.1
Posts: 995

Original Poster
Rep: Reputation: 298Reputation: 298Reputation: 298
Hi again,

Alright, will switch to nouveau to see what happens (hope it's a driver issue and not something else). If I keep getting errors, then I'll switch to the onboard graphics, and so on (to follow a proper "trial and error" procedure ).

Thanks a lot for the help!
 
Old 03-09-2012, 08:48 AM   #21
cascade9
Senior Member
 
Registered: Mar 2011
Location: Brisneyland
Distribution: Debian, aptosid
Posts: 3,718

Rep: Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903
Good luck odiseo77, hopefully you can track down the issue without too much stuffing around with drivers and pulling parts.

Quote:
Originally Posted by sycamorex View Post
That reminds me of a problem I had last year. I bought some fancy RAM (over 2000MHz). I could only set the BIOS to 1866MHz to be able to boot the machine without it freezing straight away. Fair enough. Still I noticed that while doing some CPU intensive tasks (eg. compiling a kernel with the -j5/7 flag the system would freeze. What turned out to be the problem was that my CPU only supports RAM up to 1066MHz. I hadn't known that there's any correlation between a cpu and RAM frequency.
Older x86 computers used to control the RAM from the northbridge. Newer CPUs have the memory controller intergrated into CPU.

Most DDR3 CPU memory controllers are DDR3-1066/1333. What CPU do you have? BTW, the memory doesnt actually run at 1066MHz. DDR3-1066 actually runs at 133MHz, it just uses various tricks to have a data rate of 1066MHz.

Its quite possible that you would be able to run at 1333 or 1600, even if 1866 is a problem. Even if it will only run at 1066, you can probably still lower the latency, it can make a difference with some software. Rising the latency when running at 1866 might also make the system more stable. I'd probably rather run DDR3-1600 @ 9-9-9 timings over DDR3-1866 @ 11-11-11 timings.
 
Old 03-09-2012, 09:14 AM   #22
sycamorex
LQ Veteran
 
Registered: Nov 2005
Location: London
Distribution: Slackware64-current
Posts: 5,564
Blog Entries: 1

Rep: Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026Reputation: 1026
Quote:
Originally Posted by cascade9 View Post
Older x86 computers used to control the RAM from the northbridge. Newer CPUs have the memory controller intergrated into CPU.

Most DDR3 CPU memory controllers are DDR3-1066/1333. What CPU do you have? BTW, the memory doesnt actually run at 1066MHz. DDR3-1066 actually runs at 133MHz, it just uses various tricks to have a data rate of 1066MHz.

Its quite possible that you would be able to run at 1333 or 1600, even if 1866 is a problem. Even if it will only run at 1066, you can probably still lower the latency, it can make a difference with some software. Rising the latency when running at 1866 might also make the system more stable. I'd probably rather run DDR3-1600 @ 9-9-9 timings over DDR3-1866 @ 11-11-11 timings.
I'm not at my computer now so don't remember the timings I set there. My cpu is
http://ark.intel.com/products/37147/...-GTs-Intel-QPI)

Quote:
Memory Types DDR3-800/1066
My memory:
OCZ 6GB (3x2GB) DDR3 2000MHz Gold Memory Kit CL10(10-10-10-30) 1.65V

Last edited by sycamorex; 03-09-2012 at 09:20 AM.
 
Old 03-09-2012, 10:32 AM   #23
cascade9
Senior Member
 
Registered: Mar 2011
Location: Brisneyland
Distribution: Debian, aptosid
Posts: 3,718

Rep: Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903Reputation: 903
Ahhh, an early i7/LGA 1366 setup. Yeah, technically intel speced them for DDR3-1066 max, but that was being very conservative IMO.

It should be able to to do 1333 at the very least, even without soding around with the QPI or RAM voltages. Trying to get over 1600 will normally require increasing the QPI voltage. I wouldnt push that much myself, maybe a little but not a lot. IIRC the stock voltage for i7-920 QPI is 1.1v, I wouldnt go past 1.25-1.3v. But I'd rather not push a system as hard as possible for minor gains.

I'd proably try 1333 @ 10-10-10, stock QPI voltage and stock RAM voltage, and try running the same sort of tasks where you noticed instability before. If that works, either go for 1600 @ 10-10-10, or 1333 @ 9-9-9.

BTW, its never a good idea to raise the DDR3 memory past 1.65v on LGA 1366 systems. Doing so can damage the controller. Its very, very _very_ unlikely you will damage the CPU/controller by playing with the RAM speeds though. About the worst that will happen is a boot failure due to bad settings, which you can get around by clearing the CMOS (there is normally a jumper or switch for CMOS clearing)

You can tell I'm more than happy to go into the BIOS and change settings, reboot and check the differences. I like doing that, some people cant stand it, and some people cant see the point. If you're in the 'cant be bothered/cant see why/dont like doing it' camp, I wouldnt worry to much about it.....there is difference between faster and slower RAM speeds and timings, but the place where you will notice the mosty difference is when multitasking, or doing very memory intensive tasks. For normal use, there is still a difference, buts not that big.

See here to get an idea of what I mean-

http://www.xbitlabs.com/articles/mem...rx-ddr3_5.html

If there was a few more memory intensive tasks, in particular some (real world) multitasking benchmarks, or things like video editing, the diference between DDR3-1066 and DDR3-1866 would look better. Pity that the benchmarking world is still mostly caught up in the easy 'lets do a gaming benchmark, again!' mindset....

Last edited by cascade9; 03-09-2012 at 10:40 AM.
 
Old 03-10-2012, 01:52 AM   #24
nontas.v
Member
 
Registered: Jan 2012
Posts: 32

Rep: Reputation: Disabled
Hey there, my laptop used to freeze everytime a player was playing (youtube, minitube, movie player etc), until an input form keyboard or mouse.

Try this, I hope it helps you.

Quote:
sudo gedit /etc/default/grub
In the editor window find the line wich beggins with "GRUB_CMDLINE_LINUX_DEFAULT".
After the words quiet splash add a SPACE and then type "processor.max_cstate=1" (without the quotation marks "")
So it would look like this:
Quote:
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash processor.max_cstate=1"
Save the file and then run:
Quote:
sudo update-grub
Reboot and let's hope that it's gonna work!
 
Old 03-12-2012, 04:30 PM   #25
odiseo77
Member
 
Registered: Dec 2004
Location: London, UK
Distribution: Debian Sid, OpenSUSE 13.1
Posts: 995

Original Poster
Rep: Reputation: 298Reputation: 298Reputation: 298
Quote:
Originally Posted by nontas.v View Post
Hey there, my laptop used to freeze everytime a player was playing (youtube, minitube, movie player etc), until an input form keyboard or mouse.

Try this, I hope it helps you.



In the editor window find the line wich beggins with "GRUB_CMDLINE_LINUX_DEFAULT".
After the words quiet splash add a SPACE and then type "processor.max_cstate=1" (without the quotation marks "")
So it would look like this:


Save the file and then run:


Reboot and let's hope that it's gonna work!
Thanks for the suggestion.

I'm thinking that this indeed might be a driver problem, since I switched to nouveau almost 4 days ago and the machine hasn't frozen during this time. Since the problem used to happen randomly I will have to test a little bit more to be completely sure. I'll test for about two weeks more; if it doesn't freeze, then I guess I can consider the problem solved. BTW, I checked the aptitude log, and the nvidia-glx package was upgraded from version 290.10-1 to version 295.20-1 on February 17 2012, around the time the freezes began to occur (will have to test a little bit longer, anyway).

Regards.
 
Old 03-12-2012, 04:48 PM   #26
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Hanover, Germany
Distribution: Main: Gentoo Others: What fits the task
Posts: 15,592
Blog Entries: 2

Rep: Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047
Quote:
Originally Posted by nontas.v View Post
Hey there, my laptop used to freeze everytime a player was playing (youtube, minitube, movie player etc), until an input form keyboard or mouse.

Try this, I hope it helps you.



In the editor window find the line wich beggins with "GRUB_CMDLINE_LINUX_DEFAULT".
After the words quiet splash add a SPACE and then type "processor.max_cstate=1" (without the quotation marks "")
So it would look like this:


Save the file and then run:


Reboot and let's hope that it's gonna work!
This will prevent the deeper sleep states of the CPU to kick in, causing increased power draw and heat when idling. May be good for testing, but I wouldn't do that for everyday use.
 
Old 03-13-2012, 09:48 AM   #27
nontas.v
Member
 
Registered: Jan 2012
Posts: 32

Rep: Reputation: Disabled
I know that TobiSGD, but unfortunately it was the only solution to my problem :/
 
Old 03-13-2012, 10:38 AM   #28
TobiSGD
Moderator
 
Registered: Dec 2009
Location: Hanover, Germany
Distribution: Main: Gentoo Others: What fits the task
Posts: 15,592
Blog Entries: 2

Rep: Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047Reputation: 4047
Quote:
Originally Posted by nontas.v View Post
I know that TobiSGD, but unfortunately it was the only solution to my problem :/
May be it helps to either try a newer kernel or a kernel compiled specifically for your CPU.
 
Old 03-15-2012, 11:49 PM   #29
odiseo77
Member
 
Registered: Dec 2004
Location: London, UK
Distribution: Debian Sid, OpenSUSE 13.1
Posts: 995

Original Poster
Rep: Reputation: 298Reputation: 298Reputation: 298
Well, it's been one week without freezes since I switched to nouveau (it was becoming a daily thing before I switched). I will test the machine for another week and if the system doesn't freeze during this time, I will mark the thread as solved.

Thanks to all for the help, and thanks to sycamorex for suggesting to switch to nouveau.

Cheers.
 
Old 03-18-2012, 12:45 AM   #30
cgtueno
Member
 
Registered: Jul 2003
Posts: 356
Blog Entries: 1

Rep: Reputation: 49
Hi

PSU voltage measurement is a good idea - you might have an erratic PSU.
The supply line voltages may fail to stay within acceptable levels as
the system components draw more and more current. A competent lab tech
could perform an isolated load performance test for you to eliminate
that as a cause; but given that PSUs are so cheap it is more common
to simply replace the PSU with a new one (of sufficient capacity)
(unless there are specific circumstances that preclude this as an option).

Have you cleaned the machine, removing dust and debris, and checked that
there is satisfactory airflow through the case (fans, working, air paths
not overly obstructed by cables) ?

You might be experiencing a CPU, or bridge overheat.
Check the temperature of the heatsinks to ensure that they are not
running at very high temperatures (and that your CPU fan is working ok).
If the CPU heatsink is unreasonably hot, then consider removing it,
cleaning the base of the heatsink and the top of the CPU, replacing the
thermal paste between them, and reassembling.

Running a Memtest86 for a long period is a great idea, however it doesn't
necessarily cause the CPU to heat up in the same manner that running
large numbers of processes (making up an OS running user applications)
does. This method doesn't allow for the possibility that the processing
involving the hard disks, display adapter, etc in the system might be
the cause of loading the PSU such that it fails to supply the
correct regulated voltages under higher loads.

It is an open question based on the facts presented as to whether you are
chasing an OS software, Application software, or hardware related
problem cause.

The suggestion of replacing the video card is a really good one.
I have encountered countless machines over the years that have locked up
at random times, and for no apparent reason, where the cause was
simply that the cooling fan on the display adapter had ceased to
function or was functioning in an erratic manner. When a display
adapter fails it often causes the whole system to fail because of
the close coupling of the nature of the hardware interface between
display adapter's circuitry and the CPU.

The obvious test for that scenario is to temporarily pull the display
adapter and replace it for another known working one.

So in a nut shell I suggest:
.clean the system
.inspect the system for airflow blockages
.checkout the heatsinks and fans
.check the thermal bond between the CPU and it's heatsink
.check the display adapter fan and heatsink
.change out the display adapter for a spare
.change out the PSU for a known working spare with the right capacity
and work from there.

Oh. Could it be that you are overloading a combination of PSU supply
lines. By design the majority PC switch mode PSUs have strict limitations
on the current that may be drawn from not only a single supply
line (voltage rail) but on a combination of supply lines. (It may be that
some processing event is causing multiple devices to draw sufficient
current simultaneously that is causing a PSU supply voltage regulation
problem - ie. hammering a hard disk, at same time as performing a lot
of CPU and video adapter processing - drawing lots of current - then
(say) the display adapter fan fails to keep the display processor cool, etc.

Hope that might be of some assistance.

Regards

C.
 
  


Reply

Tags
freeze


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Computer randomly freezes regardless of distro. darkrubysage Linux - General 2 12-19-2009 04:12 AM
OpenSUSE 11.0 randomly freezes VonMagnum Suse/Novell 1 10-15-2008 11:00 PM
Suse 10.1 freezes randomly preeth26 Linux - Software 20 11-13-2006 02:56 AM
computer randomly freezes randomshinichi Linux - Hardware 5 10-04-2006 04:17 AM
system freezes randomly scuffell Linux - Distributions 1 06-20-2006 06:45 PM


All times are GMT -5. The time now is 03:30 AM.

Main Menu
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration