[gentoo-user] memtest fails, but is it the RAM?

[gentoo-user] memtest fails, but is it the RAM?

[Iain Buchanan]

Hi all, slightly OT I know, but the usual excuses apply :)

I'm running memtest from a live-cd on a P4 3GHz HT desktop, with two
sticks of corsair VS512MB on an ASUS P4P800-X.

It always freezes at the start of "test 3".  The cursor keeps flashing,
and there is no display corruption, but I can't do anything but press
the reset button.

I've swapped the sticks around, used either by themselves, tried
different slots, - everything except completely different RAM.

I've never seen this behaviour with memtest before, actually, I've never
had it fail, so I don't know how it fails.  It seems a bit strange that
it fails the same way regardless of what I do - could it possibly be a
hardware/memtest incompatibility, and not actually a faulty memory
problem?

(Ultimately, I'm trying to diagnose a random reboot problem, which makes
me suspicious of the memory, but I'm not sure)

thanks heaps for the advice,
-- 
Iain Buchanan <iain at pcorp dot com dot au>

Given some of the recent threads, the interactive discussions might
need to be conducted on canvas, in the presence of a referee, while
wearing padded gloves.  ;-)
        -- Phil Hands

-- 
gentoo-user@gentoo.org mailing list

[gentoo-user] Re: memtest fails, but is it the RAM?

[James]

Iain Buchanan <iain <at> pcorp.com.au> writes:


> I've swapped the sticks around, used either by themselves, tried
> different slots, - everything except completely different RAM.

hello Iain,

Not sure this is useful, but, if you can get the system to boot, you
and look more closely at the memory specifics with the 'lshw' command.
Then if you can find the mobo book, look at the published memory 
requirements and go into the
bios, and look at the bios settings for something out of the ordinary.
If you can, swap the memory with another know good system for a few days....
Something might show up as a problem

> (Ultimately, I'm trying to diagnose a random reboot problem, which makes
> me suspicious of the memory, but I'm not sure)

I always look at the temperature as the mobo makes it available,
or checking the temperature of the hard drive with 'hddtemp /dev/<drive>'
Often power supplies will run ok can then glitch causing a
reboot. It cannot hurt to swap the power supply to see if that 
fixes the random reboots....


hth,

James



-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: memtest fails, but is it the RAM?

[Iain Buchanan]

On Fri, 2007-07-06 at 00:40 +0000, James wrote:

> Not sure this is useful, but, if you can get the system to boot, you
> and look more closely at the memory specifics with the 'lshw' command.

hey, neat command.  The system boots knoppix and windows fine, but I
don't have gentoo on it (yet :) and I don't have lshw on any live cd I
have...  It can take PC2100, PC2700 or PC3200 though.  I currently have
2x 512Mb PC3200 in it.

> Then if you can find the mobo book, look at the published memory 
> requirements and go into the
> bios, and look at the bios settings for something out of the ordinary.

it is a "jumperfree" board, with overclocking options in the bios, but
they're all turned off.

> If you can, swap the memory with another know good system for a few days....
> Something might show up as a problem

I did that.  The new RAM (another PC3200) works ok, but memtest still
fails at the same point (test 3).  This is confusing me.  the same
live-cd runs memtest on my other machine (DDR2) without fail...

The other funny thing about memtest is this:  The info it displays about
the system is a bit strange.  Sometimes it shows a CPU clock of 2999MHz,
sometimes 3000MHz; sometimes the RAM shows DDR398, sometimes DDR400.
It's always the same for one particular run of memtest, but sometimes
changes between boots.

> > (Ultimately, I'm trying to diagnose a random reboot problem, which makes
> > me suspicious of the memory, but I'm not sure)
> 
> I always look at the temperature as the mobo makes it available,
> or checking the temperature of the hard drive with 'hddtemp /dev/<drive>'

I plotted some GPU and CPU temperatures while running some games, and
they all go to a reasonable maximum and stop there.  I even turned the
case fans off, and they don't go higher.

The "random reboot" problem is now a "won't boot" problem!  I put the
original RAM back in the same slots, and now the HD's, CD's, and fans
spin up, but no display appears.  I hear a bios beep, and that's it.
Maybe it's a MBoard issue?  Maybe a video card issue?  Hmmm, I don't
want to replace the whole lot!

> Often power supplies will run ok can then glitch causing a
> reboot. It cannot hurt to swap the power supply to see if that 
> fixes the random reboots....

not able to try that... I don't have a spare.

thanks,
-- 
Iain Buchanan <iaindb at netspace dot net dot au>

The Heineken Uncertainty Principle:
	You can never be sure how many beers you had last night.

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: memtest fails, but is it the RAM?

[Dan Farrell]

On Wed, 11 Jul 2007 08:59:12 +0930
Iain Buchanan <iaindb@netspace.net.au> wrote:

> I hear a bios beep, and that's it.
> Maybe it's a MBoard issue?  Maybe a video card issue?  Hmmm, I don't
> want to replace the whole lot!

Ever hear of a POST code?  if not, look up the post codes for your
bios / motherboard and you may be able to get a head start on diagnosis
there.  

>> Often power supplies will run ok can then glitch causing a
>> reboot. It cannot hurt to swap the power supply to see if that 
>> fixes the random reboots....  

This happens sometimes with cheap power supplies, it's true.  One
temporary and inconsistant solution is to remove all unnecessary power
suckers, like hds, case fans (which evidently do nothing for you
anyway) and cdrom.  That might get it working, but not for long no
doubt.  

>the system is a bit strange.  Sometimes it shows a CPU clock of
>2999MHz,
>sometimes 3000MHz; sometimes the RAM shows DDR398, sometimes DDR400.
>It's always the same for one particular run of memtest, but sometimes
>changes between boots.

Wow, that is wierd.  Have you tried looking for BIOS updates?
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] Re: memtest fails, but is it the RAM?

[Mick]

[-- Attachment #1: Type: text/plain, Size: 2496 bytes --]

On Wednesday 11 July 2007 00:29, Iain Buchanan wrote:
> On Fri, 2007-07-06 at 00:40 +0000, James wrote:
> > Not sure this is useful, but, if you can get the system to boot, you
> > and look more closely at the memory specifics with the 'lshw' command.
>
> hey, neat command.  The system boots knoppix and windows fine, but I
> don't have gentoo on it (yet :) and I don't have lshw on any live cd I
> have...  It can take PC2100, PC2700 or PC3200 though.  I currently have
> 2x 512Mb PC3200 in it.

Use apt-get install to install the lshw package on Knoppix.  It has a small 
footprint and it should hopefully not exhaust your RAM.

> > If you can, swap the memory with another know good system for a few
> > days.... Something might show up as a problem
>
> I did that.  The new RAM (another PC3200) works ok, but memtest still
> fails at the same point (test 3).  This is confusing me.  the same
> live-cd runs memtest on my other machine (DDR2) without fail...
>
> The other funny thing about memtest is this:  The info it displays about
> the system is a bit strange.  Sometimes it shows a CPU clock of 2999MHz,
> sometimes 3000MHz; sometimes the RAM shows DDR398, sometimes DDR400.
> It's always the same for one particular run of memtest, but sometimes
> changes between boots.

Hmm, is there a BIOS firmware upgrade you could perhaps flash it with?

Can you swap around the RAM modules or remove them one at a time until you 
find the culprit?  (not sure if you tried that already).

> > > (Ultimately, I'm trying to diagnose a random reboot problem, which
> > > makes me suspicious of the memory, but I'm not sure)
> >
> > I always look at the temperature as the mobo makes it available,
> > or checking the temperature of the hard drive with 'hddtemp /dev/<drive>'
>
> I plotted some GPU and CPU temperatures while running some games, and
> they all go to a reasonable maximum and stop there.  I even turned the
> case fans off, and they don't go higher.
>
> The "random reboot" problem is now a "won't boot" problem!  I put the
> original RAM back in the same slots, and now the HD's, CD's, and fans
> spin up, but no display appears.  I hear a bios beep, and that's it.
> Maybe it's a MBoard issue?  Maybe a video card issue?  Hmmm, I don't
> want to replace the whole lot!

POST error.  Could be due to dodgy memory.  Have you tried removing the CMOS 
battery and then reflashing the BIOS with the latest firmware?
-- 
Regards,
Mick

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]