Will DDR3 memory freeze 10.04?

[J. L.]

Am 14.03.2011 23:13 schrieb scar:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
> 
> J. L. @ 03/14/2011 02:22 PM:
>> Am 14.03.2011 21:53 schrieb Larry Vaden:
>>> On Mon, Mar 14, 2011 at 3:10 PM, scar <scar at drigon.com> wrote:
>>>>
>>>> Trying desperately to figure out why a brand-new system running 10.04
>>>> 64-bit keeps freezing after 3-5 days uptime.  power supply tests good,
>>>> memtest reports no errors after going through 1 pass
>>>>
>>> Make sure you have good clean solid power and please consider running
>>> the memtest for a _MUCH LONGER_ period of time.
>>>
>> Full ack!
>>
>> I found power supply instabilities (for example 3,15 instead of the
>> specified 3,3 Volts or any other voltages) quite often to be the cause
>> for system instabilities.
>>
>> And memtest should be run "as long as possible" but at least for two
>> whole days.
>>
> 
> Thanks.
> 
> i'm reluctant to run memtest for so long since the machine is already
> deployed, but it looks i'll need to.

Even the best memory testing sw can only test a "few" / a certain amount
of all possible bit patterns in memory and some errors do not occur
regularly and only "from time to time" or under "certain circumstances".
Rising temperatures after several working hours or days should not be
neglected as possible causes. And and and ...

> so, if, after several days of memtest, errors do start showing up, how
> can i know that the memory is the culprit?  the memory is ECC so i kind
> of thought that was what ECC was for, to prevent memory errors from
> taking down the system??

Are we talking of "good" memory from for example Kingston, Corsair or
Crucial or some kind of "no-name-memory"?

And normal ECC-memory can only "correct" single-bit-failures ...

Start with one module, run the tests, add another module, run the tests,
change memory slots, run the tests ...

> is it more likely it could be the power supply?  i have one of those
> blocks that i plug the headers into and all the lights are green (except
> for one, but i contacted Antec and they said that was normal for this
> supply)

Don't know Your specific psu; but i've seen definitely more than ten
psus that caused problems for different reasons: Rising temperatures,
"cheap" capacitors (on the maiboard for example too), unstable or too
low or too high voltages especially but not only under (high) load
conditions and so on ...

Try to check at least the voltages with a circuit analyser or multimeter
or perhaps simply try to check another psu.

Good luck!

J. L.