Re: memory corruption under heavy load?

Kevin M Bealer (kmb203@psu.edu)
Wed, 3 Apr 1996 23:51:33 -0500 (EST)


On Tue, 2 Apr 1996, Marek Michalkiewicz wrote:

> tim middlekoop:
> >
> > I had problems along simular lines, disabling the external cache worked
> > also, could it be that the kernel is having problems with external cache,
> > or just slow memory.
>
> I'm getting a lot of mail about slow memory :-(. Could someone finally
> explain this to me: it works fine on the same machine under 1.2.13 and
> 1.3.58. Why??? Could the SIMMs fix themselves when running a stable
> kernel? :-)
>
> Marek
>

This has been covered sporadically, some ram defects only occur under very
obscure conditions, certain bits in certain places, ram which is borderline
too slow, or other conditions. Any minor change in software could radically
alter the low level usage of the hardware.

So it seems the obstacles can be avoided by tweaking, twiddling and
frobnicating software, but the only test of whether it is hardware is if it
doesn't work and changes of hardware fix it.

BTW: I'm only repeating what others have said. It sounds ok to me, but I
don't actually _know_.

__kmb203@psu.edu_____________________________Debian/GNU__Linux__1.3.77___
So you think you know the real meaning of fear?
Yeah, you think you do know, but I doubt it.
When you sit in a shelter with bombs falling all over.
And the houses around you are burning like torches.
I agree that you experience horror and fright
For such moments are dreadful, for as long as they last,
But the all-clear sounds--then it's okay-- | -- Ilya Selvinskiy
You take a deep breath, the stress has passed by. | (Taken from "The Sum
But real fear is a stone deep down in your chest. | of All Fears" by
You hear me? A stone. That's what it is, no more. | Tom Clancy pg. 182.)