Defective L2 Cache

Lennart Sorensen lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org
Tue Jan 18 14:58:07 UTC 2005


On Tue, Jan 18, 2005 at 01:55:27AM -0500, Logan Rathbone wrote:
> Hi guys.  I've had a bit of a problem for quite a while, but it has only
> become very annoying as of late.
> 
> I have a problem with my memory, and I'm 99% sure it's my L2 cache.  It
> all started when I upgraded my motherboard last year.  I was on a bit of
> a budget, so I opted for a rather cheap board (I can't even think of the
> brand off hand...) with an AMD K7 2.2 GHz processor (again not sure
> exactly what its "brand" name is.. but those are its tech specs).  I
> also upgraded my RAM from 128 to 512, which is already beginning to show
> its age ;-)
> 
> Anyway, here's the dealio.  I noticed that after I left my computer
> running for a while, compiling kernels and the like, my compilations
> would suddenly stop with a SIGSEGV, an internal compiler error.  At
> first it wasn't a big deal; I just had to start from where I left off.
> 
> But still, I decided to run memtest86.  I'm a fairly patient person; but
> after it had been running for over 2 days without reporting a single
> error, I had to stop it.  There's work to do, y'know.
> 
> But now, a year later, I run into problems that reach as far as the user
> level.  If I leave my computer on for about 24 hours or so, programs
> start randomly crashing (and I'm talking EVERYTHING, from VIM to Mozilla
> to irssi).
> 
> Reluctant to try memtest86 again, I found a very helpful page about
> deducing memory problems, and finding out where they lie.  I think I
> narrowed it down to the L2 cache -- it's a 256-meg btw -- I disabled the
> L2 cache, and while the system was slow as hell, I didn't get any
> internal compiler errors, and everything was otherwise fine.
> 
> I used to have Windows XP a while ago, but since I upgraded my hard
> drive (previous thread on here somewhere...) I haven't gotten around to
> re-installing it.  But when it was there, it seemed to run without
> problems.  But that's not exactly a fair comparison, as I never compile
> anything on Windows, and don't really use it all that often, just for my
> family.
> 
> So my question is... what do I do to fix this?  Does the part have to be
> replaced?  And why would it not cause any problems whatsoever on
> Windows, what sets it apart in that respect?
> 
> Sorry for the long-winded post, but I hope someone here can help ;-)

Well a few potential causes:

The L2 cache might be defective, in which case you either run with L2
off, or you replace the CPU (since that is where the L2 cache is).

Maybe your power supply is garbage (quite likely if you bought a cheap
no name power supply, which I would suspect if your motherboard is of
unknown origin too).  I have seen a K6 450 in the past that a friend had
which would crash regularly and corrupt the filesystem (it was mostly
running win98 at the time).  Eventually my friend happened to look at
the power supply monitor in the bios (asus boards have had that for
years) and noticed the 3.3V line was showing 2.9V.  He went and got a
quality name brand power supply, and amazingly the system stopped
crashing, the filesystem didn't corrupt anymore, and it just ran.  Even
win98 stopped crashing (except when running badly written games, but
what can you do about that).

Quality matters.  You usually don't get what you don't pay for (too bad
you get what you pay for isn't always true).

A crappy motherboard (it contains the voltage regulators) or power supply
can make any cpu misbehave even if the cpu is ok.  I have seen cpus fail
too though so it isn't imposible that the cpu is bad.

Lennart Sorensen
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml





More information about the Legacy mailing list