Hi.
I've got some feedback for you, as I've just had to deal with a bad
DIMM myself in a Mac Cube G4.
On Fri, Jan 25, 2002 at 12:12:56PM +0200, Tuomas Kuosmanen wrote:
> On Fri, 2002-01-25 at 07:06, Michel Lanners wrote:
> > On 25 Jan, this message from Russell Hires echoed through cyberspace:
> > >
> > > Just so I can be sure it's not my RAM causing my kernel oopses,
> > > is there a utility for testing RAM in Linux? I know that there
> > > is one in the x86 world that can test memory, but in the powerpc
> > > world...???
> >
> > None that I know off (and this has been asked a few times before,
> > without answer...).
>
> Check out
>
> http://www.qcc.sk.ca/~charlesc/software/memtester/
>
> This one runs on ppc. I cannot tell if it is good or bad, since I dont
> know how a memory test should work to be "good", but it prints nice
> stats to STDOUT, hogs the CPU and allocates a lot of RAM. I guess it
> works ;-)
I had a DIMM go bad about two months after I installed it, which I
noticed because something kept corrupting my available cache for apt
and dselect. I had 3 512MB PC133 CAS2 DIMMS in there (hey, why not,
they were only $42 apiece when I bought them), all from the same
batch, so I was afraid that all three were bad and / or incompatible
with the Cube, and I wasn't really looking forward to debugging that
whole mess.
Anyway, the first article in this thread prodded me into finally
dealing with my poor little Cube, so I did a little searching in the
list archives, and came up with the above link to memtester. I built
it and started it up, and it found errors almost immediately, and only
with one of the DIMMs. It took me less than an hour to find the bad
DIMM. After buying a replacement DIMM today (sadly, it cost me $80 and
is only CAS3), am running an extended burn-in test, in an effort to
ensure that I don't have to go through this rigamarole again.
The program uses malloc() and mlock() to grab the largest available
single chunk of memory it can, which in the case of my 1.5GB system
comes to about 800MB. It was able to malloc() 1275 MiB, but only able
to mlock() ~768MB, which probably doesn't even span all three
DIMMS. It seems pretty clear to me, then, that it can't be a
comprehensive test, although it's very thorough with what it does do
(it takes ~2 hours on my Cube to run one pass with 1 GB of RAM
installed, and ~3 1/2 hours to test the full 1.5 GB).
A related, if off-topic question: does anyone know if the Cube's
hardware groks CAS lines (does any Mac hardware?), and if so, if
mixing CAS2 and CAS2 RAM is going to cause difficulty? I assume that
they'll all fall back to CAS3, but I'm not even sure that having CAS2
RAM would have made any difference to performance in the first place.
Obviously, I don't care enough to try to benchmark it, but I'd like to
know if any you have the information handy.
> > If it's SDRAM, stick it in a i386 and test there?
>
> That too.
I'm going to use memtest86 in my PC to see if my PC hates the bad
DIMM, and then maybe rotate the three DIMMs that are in the Cube, just
to be sure. I'll post a followup with the results.
> I wonder how hard it would be to port memtest86 to ppc, it is a neat
> one since it is basically a linux kernel that just runs the memory
> test. Nice way to minimize the reserved memory size.. But I guess
> it also writes its output directly to the VGA memory, which would be
> a problem for PPC I guess.
Several important glue files in Memtest86 are x86 assembler code (~2k
lines' worth of assembler), and it hauls around a precompiled kernel
of its own that's x86 as well for use on boot floppies. I think you
mean "rewrite" instead of "port". ;)
I'm slightly appalled that better RAM-testing software isn't
available, free or otherwise, given how much more, ah, "delicate" RAM
is these days (which doesn't even outweigh the fact that RAM seems to
be included in cereal boxes these days, but is still a pain to deal
with).
Forrest
--
. . . the self-reflecting image of a narcotized mind . . .
ozymandias G desiderata ogd@aoaioxxysz.net desperate, deathless
(415)823-6356 http://www.pushby.com/forrest/ ::AOAIOXXYSZ::
Attachment:
pgp3UggQw0usX.pgp
Description: PGP signature