James Mastros wrote:
>
> On Fri, 29 May 1998, Andreas Haumer wrote:
> > 9.) MB ASUS P/I-P55T2P4, 256MB RAM, Intel i586-200 MMX, 2xAHA2940UW,
> > 2GB SCSI + 16GB UW-SCSI RAID5 (5xIBM DCAS-34330), 120MB Swap,
> > Linux-2.0.34pre16, gcc-2.7.3.2
> > (same system as 6., but with Intel CPU)
> > -> NO Crash after 3 hours (about 30000 Processes), and still
> > running...
> I also noticed that all of your tests are under 2.1.33 or .34 -- I'm
No, 2.0.34pre1[56]
^^^
> starting one under 2.1.104-pre1 now.
>
> 10.) MB Unknown, 32MB RAM, Intel i586-166 (stepping 12), Triton VX chipset,
> 32MB swap, gcc-2.8.0.
> -> NO Crash after about 5 min <G>, still running (and under active
> use, though I'm killing rc5 now) (anybody up for
> distributed-crashme <G>).
>
> > It seems, "crashme" triggers some problem within the AMD CPU, maybe
> > some illegal code halts or crashes the CPU or something like that.
> > (Doesn't this sound familiar? We all remember the problems AMD CPU's
> > had about a year ago!)
> Hmm... but the different systems crashed at very different points, whereas
> if all of them were crashing on some piricular instruction, you would expect
> the time of the crash to scale with the processor speed. (Even though we
> are using a psudo-random string, we are always initing with the same value,
> so we should get the same string, yes?)
Yep, the "seed" used was always the same, so the stream of random
numbers
should have been the same, too. But I have to agree that the chances
to
find one single, independent data string which crashes the K6 are
probably
quite low.
The fault might very well be dependent on many different things (as a
CPU
usually is a quite complex piece of silicon... :-))
But: The crashes can be reproduced, and they can be reproduced quite
fast
(it usually lasts less than 15 minutes), so it's worth to run some
tests
and try to find a pattern.
Anyway, it looks like I would have to avoid the K6 in Linux servers in
the near future... :-(
- andreas
PS: "crashme" currently is still running on test #9: I'm now at 4
hours
and 50 minutes, with more than 50000 processes created.
-- Andreas Haumer | email: andreas@xss.co.at | PGP key available *x Software + Systeme | phone: +43.1.6001508 | on request. Buchengasse 67/8 | +43.664.3004449 | A-1100 Vienna, Austria | fax: +43.1.6001507 |- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu