Re: my video card is possessed

2009-04-13 Thread Craig Falconer

Aidan Gauland wrote, On 12/04/09 09:26:
...
This is looking more and more likely... to me, a programmer who avoids 
dealing with hardware.


So, any recommendations for an AGP video card?


They're hard to find new these days.  Look for a cheap one off trademe, 
or stump up for a new board/CPU/ram then migrate the old board to be a 
server with a PCI video card.



--
Craig Falconer



Re: my video card is possessed

2009-04-11 Thread Adrian Mageanu


On Sat, 2009-04-11 at 16:56 +1200, Adrian Mageanu wrote:
 Maybe you've done this already, but just asking, did you run
 sensors-detect? This will make your OS aware of the available
 temperature sensors in your system.
 
 Adrian

  
  DISPLAY=:0 nvidia-settings -q GPUCoreTemp
 

My mistake, you don't need to run sensors-detect if you run the command
above to get only the temperature of the video card



Re: my video card is possessed

2009-04-11 Thread Aidan Gauland
If I run the GUI, it tells me that I do not appear to be using the NVIDIA X 
driver, and then only gives me options for the GUI itself, but not the card.


Nick Rout wrote:

well does nvidia-settings run? it has a gui too, mine shows gpu core
temp, but its a recent card.


Re: my video card is possessed

2009-04-11 Thread Aidan Gauland
I can't reassure you *that* far.  But the only components that I have touched 
before this problem started to occur are the drives.


Christopher Sawtell wrote:

Can you re-assure us that it is completely impossible that your
machine could have suffered damage from static electricity?

That is to say, you have _always_ used an anti-static wrist strap when
touching _any_ part of your computer?


Re: my video card is possessed

2009-04-11 Thread Nick Rout
well does nvidia-settings run? it has a gui too, mine shows gpu core
temp, but its a recent card.

On Sun, Apr 12, 2009 at 4:39 AM, Aidan Gauland
wgsil...@no8wireless.co.nz wrote:
 I already tried just
 nvidia-settings -q GPUCoreTemp
 and it gave the same output.
 BTW $DISPLAY=:0.0

  -Aidan

 Euan Clark wrote:

 Your DISPLAY my be running on :1  - try it without the DISPLAY=:0



Re: my video card is possessed

2009-04-11 Thread Euan Clark
Have you had any drive issues? Check for IDE cables being dodgy and/or 
shorting to the case - I had  a recent experiance similar to yours where 
this turned out  to be the issue.


Aidan Gauland wrote:
I can't reassure you *that* far.  But the only components that I have 
touched before this problem started to occur are the drives.


Christopher Sawtell wrote:

Can you re-assure us that it is completely impossible that your
machine could have suffered damage from static electricity?

That is to say, you have _always_ used an anti-static wrist strap when
touching _any_ part of your computer?






Re: my video card is possessed

2009-04-11 Thread Aidan Gauland

Nope, my drives have never let me down.  And the cables look fine.

Euan Clark wrote:
Have you had any drive issues? Check for IDE cables being dodgy and/or 
shorting to the case - I had  a recent experiance similar to yours where 
this turned out  to be the issue.


Re: my video card is possessed

2009-04-11 Thread Aidan Gauland

Aidan Gauland wrote:

Right, here's what I've ruled out (and why)...

[snip]

Let's add RAM to that list.  The memory test passed (I only ran it for 10 
hours; I was too impatient to wait for the rest to finish).



So that leaves...

[snip]

* The power supply is not delivering enough power.
Uhhh, right, I'll have to find a working power supply to swap with my 
current one, if I'm going to rule this one out.  Unless someone has a 
better idea.


* The graphics card is ed, and needs to be replaced.
This is looking more and more likely... to me, a programmer who avoids 
dealing with hardware.


So, any recommendations for an AGP video card?

 -Aidan


Re: my video card is possessed

2009-04-11 Thread Aidan Gauland

Aidan Gauland wrote:

So, any recommendations for an AGP video card?


I suppose I should have said what I need to do with it: I play with Blender a 
bit, but I don't do anything complex with it; and I use Pipmak 
http://pipmak.sourceforge.net/ which needs 3D hardware acceleration for 
displaying cubic panoramas.  But I don't need a card for hard-core gaming.


So, is there anything I should avoid or look out for when shopping for a 
replacement card?


Thanks,
Aidan


Re: my video card is possessed

2009-04-10 Thread Euan Clark

What temperatures were you seeing?

Divide and conquer time. System lockups may be occurring because of 
another component going awry as the system comes up to normal working 
temperature. Coud be a marginal-seating issue.
Check the seating of the PCI cards and all power connectors. Another of 
the usual suspects is the RAM - are they seated properly? If so run a 
memtest using a liveCD, e.g. Gentoo-minimal has one.
Running up a livecd may also give you some indication of it being a 
hardware issue or related to the install you have on disk.
Also eliminate that the motherboard isn't shorting to the case in some 
way, e.g. loose screw floating round, dust etc.


Nvidia split their drivers at some point into nvidia-drivers and 
legacy-drivers. I'm unsure at what point this occurred and which one is 
for your card.

hmmn - AGP Bus - can we see the output of sudo lspci and cat /proc/cpuinfo

Whats the history on this box and whats on it? Any chance it has been 
knocked around? Any out-of-package-management / compiled-from-source 
installs of drivers? Recent BIOS upgrades?


- Euan



Aidan Gauland wrote:
 Ok, I've just tried the nv driver with a huge desk fan, on high, 
pointed at the graphics card heat-sinks.  I also tried detaching my 
system fan from the case and doing the same thing with it.   No 
diffreence; it still freezes after a little while.  It's a low-end 
video card (see attached file for full information), and has no fan to 
fail.  Is there a way to test the card for heat damage (or any kind of 
subtle damage)?


 And, John, how could the motherboard battery affect the video?

Thanks,
Aidan

John Mallett wrote:

On Sunday 05 April 2009 17:02:47 Steve Holdoway wrote:

Sounds to me like the fan's failed...
If it is an older computer doing strange things it might be the 
battery on the motherboard. It wouldn't hurt to replace it.




Re: my video card is possessed

2009-04-10 Thread Aidan Gauland

Euan Clark wrote:

What temperatures were you seeing?


My video card doesn't seem to have temperature sensors that I can see from my 
OS.  But I don't know how to access them, even if there are any that I can 
access. I forgot to mention in my last post that I put my fingers on the GPU 
heat-sinks after the display forze, and it did not feel hot, just luke-warm.


Divide and conquer time. System lockups may be occurring because of 
another component going awry as the system comes up to normal working 
temperature. Coud be a marginal-seating issue.
Check the seating of the PCI cards and all power connectors. Another of 
the usual suspects is the RAM - are they seated properly? If so run a 
memtest using a liveCD, e.g. Gentoo-minimal has one.
Running up a livecd may also give you some indication of it being a 
hardware issue or related to the install you have on disk.
Also eliminate that the motherboard isn't shorting to the case in some 
way, e.g. loose screw floating round, dust etc.


I'll run a memtest tonight.  Everything seems to be in there pretty snuggly. 
I took the video card out a couple of days ago to write down any helpful 
looking numbers that were on it.  I'll check for any shorts, but I've looked 
closely at the entire inside of the box, and did not see anything suspicious. 
 I'll try John's suggestion, and replace the motherboard battery.


Nvidia split their drivers at some point into nvidia-drivers and 
legacy-drivers. I'm unsure at what point this occurred and which one is 
for your card.

hmmn - AGP Bus - can we see the output of sudo lspci and cat /proc/cpuinfo


Ok, see attached files.

Whats the history on this box and whats on it? Any chance it has been 
knocked around? Any out-of-package-management / compiled-from-source 
installs of drivers? Recent BIOS upgrades?


I bought my box second-hand through SellMeFree.co.nz.  I don't know its 
history before that.  That was over a year ago, and it's been running 
perfectly until now.


Considering that Live CDs and a fresh install of Debian freeze, I don't think 
it's a software-side problem (and no, I'm only using system components from 
Debain).  And I've never upgraded the BIOS since I bought the thing.


Thanks,
Aidan


cpuinfo.gz
Description: application/gzip


lspci.out.gz
Description: application/gzip


Re: my video card is possessed

2009-04-10 Thread Volker Kuhlmann
On Fri 10 Apr 2009 16:39:06 NZST +1200, Aidan Gauland wrote:

 for full information), and has no fan to fail.  Is there a way to test the 
 card for heat damage (or any kind of subtle damage)?

No there isn't. Not for heat damage, and not for anti-static damage. If
the card doesn't have a fan, and doesn't have heats sinks clogged with
dust, then it doesn't have heat damage. If it is subtly damaged, the
symptoms you get are unreliability. Same as for old age.

  And, John, how could the motherboard battery affect the video?

Not likely, but it's always worth a shot to at least clear the cmos and
reload factory/safe/whatertheycalledit defaults. Replace the battery
only if it has  2.9V or so, but do rub the clip contacts off and wipe
the battery hard with a cotton cloth, then insert it without(!) leaving
fingerprints (or lint). Skin grease corrodes metal surfaces over time.
The real reason you want that battery to work is so that it doesn't gain
funny settings while you power the box off (and to keep the time if you
don't use ntp).

Tracking down unreliability issues is difficult. Faults that only occur
sometimes are hard to find. Pretty much all has been said. power
supplies can lose uumph, unless you have an electronics workshop, swap
it with another one and see if problems disappear.

As you tried different distros I agree that it is most likely a hardware
and not a software fault.

Volker

-- 
Volker Kuhlmann is list0570 with the domain in header
http://volker.dnsalias.net/ Please do not CC list postings to me.


Re: my video card is possessed

2009-04-10 Thread Nick Rout
On Sat, Apr 11, 2009 at 9:21 AM, Aidan Gauland
wgsil...@no8wireless.co.nz wrote:
 Euan Clark wrote:

 What temperatures were you seeing?

 My video card doesn't seem to have temperature sensors that I can see from
 my OS.  But I don't know how to access them, even if there are any that I
 can access. I forgot to mention in my last post that I put my fingers on the
 GPU heat-sinks after the display forze, and it did not feel hot, just
 luke-warm.

DISPLAY=:0 nvidia-settings -q GPUCoreTemp


Re: my video card is possessed

2009-04-10 Thread Aidan Gauland

Nick Rout wrote:

DISPLAY=:0 nvidia-settings -q GPUCoreTemp


Odd, the output of that is just an empty line.


Re: my video card is possessed

2009-04-10 Thread Aidan Gauland

Nick Rout wrote:

DISPLAY=:0 nvidia-settings -q GPUCoreTemp


Odd, the output of that is just an empty line.  Does that mean that my card 
has no sensors that I can use?


Re: my video card is possessed

2009-04-10 Thread Euan Clark

Aidan Gauland wrote:

Nick Rout wrote:

DISPLAY=:0 nvidia-settings -q GPUCoreTemp


Odd, the output of that is just an empty line.


Your DISPLAY my be running on :1  - try it without the DISPLAY=:0



Re: my video card is possessed

2009-04-10 Thread Aidan Gauland

I already tried just
nvidia-settings -q GPUCoreTemp
and it gave the same output.
BTW $DISPLAY=:0.0

 -Aidan

Euan Clark wrote:

Your DISPLAY my be running on :1  - try it without the DISPLAY=:0


Re: my video card is possessed

2009-04-10 Thread Adrian Mageanu
Maybe you've done this already, but just asking, did you run
sensors-detect? This will make your OS aware of the available
temperature sensors in your system.

Adrian


On Sat, 2009-04-11 at 10:58 +1200, Nick Rout wrote:
 On Sat, Apr 11, 2009 at 9:21 AM, Aidan Gauland
 wgsil...@no8wireless.co.nz wrote:
  Euan Clark wrote:
 
  What temperatures were you seeing?
 
  My video card doesn't seem to have temperature sensors that I can see from
  my OS.  But I don't know how to access them, even if there are any that I
  can access. I forgot to mention in my last post that I put my fingers on the
  GPU heat-sinks after the display forze, and it did not feel hot, just
  luke-warm.
 
 DISPLAY=:0 nvidia-settings -q GPUCoreTemp



Re: my video card is possessed

2009-04-10 Thread Aidan Gauland

Right, here's what I've ruled out (and why)...

* Heat damage
There's no GPU fan, and the heat-sinks are clean.

* Overheating
I've run my machine with a giant fan blowing, on high, across the GPU 
heat-sinks.

* Motherboard battery
It's above 3 volts, so it still has a good charge.  And I took it out for a 
few minutes, then put it back in, to reset the cmos.


* Misbehaving system software
I've tried several different distros.  They all freeze.

So that leaves...

* RAM
I'll run a memtest tonight.

* The power supply is not delivering enough power.
Uhhh, right, I'll have to find a working power supply to swap with my current 
one, if I'm going to rule this one out.  Unless someone has a better idea.


* The graphics card is ed, and needs to be replaced.
This is looking more and more likely... to me, a programmer who avoids dealing 
with hardware.


When the hardware goes wonky, the universe in which the software exists ceases 
to obey its normal, somewhat sane and logical laws.  It seems as if I'm 
working in a universe in which two may sometimes equal one.


Thanks for all the help... so far.  I'm not done yet.
 -Aidan


Re: my video card is possessed

2009-04-10 Thread Steve Holdoway
On Sun, 12 Apr 2009 04:33:55 +1200
Aidan Gauland wgsil...@no8wireless.co.nz wrote:

 Nick Rout wrote:
  DISPLAY=:0 nvidia-settings -q GPUCoreTemp
 
 Odd, the output of that is just an empty line.  Does that mean that my card 
 has no sensors that I can use?

Probably yes. It's an old card and you're probably using the legacy nvidia 
drivers. This lappie uses them with GeForce4 460 Go graphics, which is ancient.

Steve
-- 
Steve Holdoway st...@greengecko.co.nz


Re: my video card is possessed

2009-04-10 Thread Aidan Gauland

Steve Holdoway wrote:

On Sun, 12 Apr 2009 04:33:55 +1200
Aidan Gauland wgsil...@no8wireless.co.nz wrote:
Odd, the output of that is just an empty line.  Does that mean that my card 
has no sensors that I can use?


Probably yes. It's an old card and you're probably using the legacy nvidia 
drivers. This lappie uses them with GeForce4 460 Go graphics, which is ancient.


Er, no, I'm not using the legacy drivers.  I'm stuck with vesa.  Otherwise my 
system freezes.


Re: my video card is possessed

2009-04-10 Thread Steve Holdoway
On Sun, 12 Apr 2009 05:02:56 +1200
Aidan Gauland wgsil...@no8wireless.co.nz wrote:

 Steve Holdoway wrote:
  On Sun, 12 Apr 2009 04:33:55 +1200
  Aidan Gauland wgsil...@no8wireless.co.nz wrote:
  Odd, the output of that is just an empty line.  Does that mean that my 
  card 
  has no sensors that I can use?
  
  Probably yes. It's an old card and you're probably using the legacy nvidia 
  drivers. This lappie uses them with GeForce4 460 Go graphics, which is 
  ancient.
 
 Er, no, I'm not using the legacy drivers.  I'm stuck with vesa.  Otherwise my 
 system freezes.

Ah, you'll get no info off the card without running the nvidia drivers. I've 
got a spare old pcie nvidia card you can try, but nothing in any other format 
any more ):

Steve
-- 
Steve Holdoway st...@greengecko.co.nz


Re: my video card is possessed

2009-04-10 Thread Christopher Sawtell
Can you re-assure us that it is completely impossible that your
machine could have suffered damage from static electricity?

That is to say, you have _always_ used an anti-static wrist strap when
touching _any_ part of your computer?

-- 
Sincerely etc.
Christopher Sawtell


Re: my video card is possessed

2009-04-09 Thread Aidan Gauland
 Ok, I've just tried the nv driver with a huge desk fan, on high, pointed at 
the graphics card heat-sinks.  I also tried detaching my system fan from the 
case and doing the same thing with it.   No diffreence; it still freezes after 
a little while.  It's a low-end video card (see attached file for full 
information), and has no fan to fail.  Is there a way to test the card for 
heat damage (or any kind of subtle damage)?


 And, John, how could the motherboard battery affect the video?

Thanks,
Aidan

John Mallett wrote:

On Sunday 05 April 2009 17:02:47 Steve Holdoway wrote:

Sounds to me like the fan's failed...
If it is an older computer doing strange things it might be the battery on the 
motherboard. It wouldn't hurt to replace it.


gfxcardInfo.txt.gz
Description: application/gzip


Re: my video card is possessed

2009-04-07 Thread Craig Falconer

Euan Clark wrote, On 07/04/09 17:27:
Quick confirmation of heat issues before spending money - point a 
deskfan or hairdryer (on cold) into affected areas in your open case and 


Why cold?  Hot would help make the problem reoccur... helps test the 
positive case as well  :)


I've generally left the slot beside the video card empty, and even leave 
the slot cover off to allow the air to move straight from front to back.



--
Craig Falconer



Re: my video card is possessed

2009-04-06 Thread Aidan Gauland
That sounds like my situation (the bit about passive cooling).  I'll try that. 
 Did you fix the fan in place, or just let it dangle?  There's a place in my 
box for a system fan, but there's not one there, so I'll get one.  Also, 
could you tell me a little bit more about your graphics hardware?


Thanks,
Aidan

Euan Clark wrote:
I've noticed increased heat issues on both my GPUs  on both XP  Gentoo 
- I think latest nvidia drivers push the cards harder (PhysX?)  - they 
were passive cooling only so I dropped another fan in the box pointed 
across the cards and the issues have gone.




Re: my video card is possessed

2009-04-06 Thread John Mallett
On Sunday 05 April 2009 17:02:47 Steve Holdoway wrote:
 On Sun, 05 Apr 2009 16:40:13 +1200
 Aidan Gauland wgsil...@ihug.co.nz wrote:

 [snip]
 Sounds to me like the fan's failed...
If it is an older computer doing strange things it might be the battery on the 
motherboard. It wouldn't hurt to replace it.

 


Re: my video card is possessed

2009-04-06 Thread Euan Clark


Core 2 plus 2 x Nvidia Geforce 8600 driving 2 monitors both as SLI 
(gaming/XP) and otherwise (work/Gentoo).


I have a fairly good quality full tower case with two input fans sucking 
in air across two drive arrays and extract with the main system fan. The 
trouble was that airflow wasn't really effective (a dead spot) over the 
GPU heatsinks. With latest Nvidia drivers packaging PhysX I started 
exhibiting heat issues in GPUs and nearby disk controller. I added a 
cheap free floating fan that just sits vertically on the floor of the 
case pointing across the GPU heatsinks towards some unused PCI slots I'd 
uncovered. Will fix the fan in place when I get a chance :-)  Fan was a 
$10 120mm fan that could piggyback into the power feed for the disks.
Cards went from around 70-90 degrees underload down to around a more 
acceptable 45-50 and problems disappeared.


You could also disable PhysX if it's available as an option. If the 
system fan in your box isin't populated then heat is a strong suspect. A 
lot of off-the-shelf systems run just inside the margins to save cost.


Quick confirmation of heat issues before spending money - point a 
deskfan or hairdryer (on cold) into affected areas in your open case and 
monitor temperature differences  with Speedfan or Gkrellm under load.




Aidan Gauland wrote:
That sounds like my situation (the bit about passive cooling).  I'll 
try that.  Did you fix the fan in place, or just let it dangle?  
There's a place in my box for a system fan, but there's not one 
there, so I'll get one.  Also, could you tell me a little bit more 
about your graphics hardware?


Thanks,
Aidan

Euan Clark wrote:
I've noticed increased heat issues on both my GPUs  on both XP  
Gentoo - I think latest nvidia drivers push the cards harder 
(PhysX?)  - they were passive cooling only so I dropped another fan 
in the box pointed across the cards and the issues have gone.







my video card is possessed

2009-04-04 Thread Aidan Gauland

Hi,

 I upgraded from Debian etch to lenny over a week ago.  Shortly thereafter, 
my system's display started freezing after a few minutes of use: the only 
thing that does not get affected is the cursor in X; I couldn't even get to a 
virtual console.  But I could shut it down by pressing the power button (which 
seems to be equivalent to running shutdown -h now).  I uninstalled all xorg 
drivers, except for vesa; this seems to keep my system stable, but I still get 
the freezing sometimes when I shutdown from within X.  (Quick summary of that 
oddness: *everything* in X freezes, except for the cursor; and I can't do 
*anything*, not even get to a virtual console, except for shut down (*NOT* a 
hard reset) by pressing the power button on my box.)


 Now things start resembling Steven King's 1408.  I installed Antix Mepis, 
because it describes itself as a distro for old hardware, and I assumed that I 
had found a bug in the nv xorg driver (I'm using an old (it's for AGP) nVidia 
card) that would be patched in a distro for old hardware.  After going through 
the installation process and configuring my new system, it froze, just as 
before, while Synaptic was getting the repository indexes.


 So I went back to etch.  A *FRESH INSTALLATION* of etch.  And it froze... 
while Synaptic was getting the repository induces.


 All of this was much, much more painful than it sounds.  I'm *now* using 
Debian lenny with the vesa driver.  I *was* using Debian etch with nVidia's 
non-free driver, and it was very stable.


 Does anybody have any advice?  Please?  I'm now thinking (after etch froze) 
that something has happened to my video card.  And I'm a little worried that 
it will get worse.


 Oh, yeah, and a while ago when I was using etch, some characters (it looks 
like it's always the same ones) on the screen during boot time (when it's in 
plain vga mode) would turn into a vertical bar character (this - '|').  It 
still happens, and to more characters than when it started (i.e. it's gotten 
worse).  So, again, taking all of this into consideration, I think my video 
card is on its way out, and needs to be replaced.


Thanks,
Aidan




Re: my video card is possessed

2009-04-04 Thread Steve Holdoway
On Sun, 05 Apr 2009 16:40:13 +1200
Aidan Gauland wgsil...@ihug.co.nz wrote:

[snip]
Sounds to me like the fan's failed...


-- 
Steve Holdoway st...@greengecko.co.nz


Re: my video card is possessed

2009-04-04 Thread Euan Clark
I've noticed increased heat issues on both my GPUs  on both XP  Gentoo 
- I think latest nvidia drivers push the cards harder (PhysX?)  - they 
were passive cooling only so I dropped another fan in the box pointed 
across the cards and the issues have gone.



Aidan Gauland wrote:

Hi,

 I upgraded from Debian etch to lenny over a week ago.  Shortly 
thereafter, my system's display started freezing after a few minutes 
of use: the only thing that does not get affected is the cursor in X; 
I couldn't even get to a virtual console.  But I could shut it down by 
pressing the power button (which seems to be equivalent to running 
shutdown -h now).  I uninstalled all xorg drivers, except for vesa; 
this seems to keep my system stable, but I still get the freezing 
sometimes when I shutdown from within X.  (Quick summary of that 
oddness: *everything* in X freezes, except for the cursor; and I can't 
do *anything*, not even get to a virtual console, except for shut down 
(*NOT* a hard reset) by pressing the power button on my box.)


 Now things start resembling Steven King's 1408.  I installed Antix 
Mepis, because it describes itself as a distro for old hardware, and I 
assumed that I had found a bug in the nv xorg driver (I'm using an old 
(it's for AGP) nVidia card) that would be patched in a distro for old 
hardware.  After going through the installation process and 
configuring my new system, it froze, just as before, while Synaptic 
was getting the repository indexes.


 So I went back to etch.  A *FRESH INSTALLATION* of etch.  And it 
froze... while Synaptic was getting the repository induces.


 All of this was much, much more painful than it sounds.  I'm *now* 
using Debian lenny with the vesa driver.  I *was* using Debian etch 
with nVidia's non-free driver, and it was very stable.


 Does anybody have any advice?  Please?  I'm now thinking (after etch 
froze) that something has happened to my video card.  And I'm a little 
worried that it will get worse.


 Oh, yeah, and a while ago when I was using etch, some characters (it 
looks like it's always the same ones) on the screen during boot time 
(when it's in plain vga mode) would turn into a vertical bar character 
(this - '|').  It still happens, and to more characters than when it 
started (i.e. it's gotten worse).  So, again, taking all of this into 
consideration, I think my video card is on its way out, and needs to 
be replaced.


Thanks,
Aidan