From: Test Account Fachschaft E-Technik (fset@guug.de)
Date: 04/28/93


From: fset@guug.de (Test Account Fachschaft E-Technik)
Subject: Re: 0.99.8 kernel panic ...
Date: Wed, 28 Apr 1993 17:52:19 GMT

eilhard@uni-paderborn.de (Oliver Eilhard) writes:

>I had the same problem with my machine, but I did not get
>a SYSTEM HALTED message. The system just stopped.
>I had a look at the BIOS-Setup of my machine, and there was
>a mechanism called "Decoupled refresh". As I disabled this mechanism,
>I had no problems until now. I did not tested it for very long, but
>normally Linux halted just after a few minutes. And with disabling
>the mechanism, the system worked for over half an hour.
>If anybody knows what "Decoupled refresh" really is, please
>mail me.

>By the way, of course this is no Linux-dependend problem. My machine
>stopped under various OS, but it's a fact that Linux just needed
>a few minutes to stop while OS/2 needed about 15 minutes. This proves,
>that Linux makes intensive use of 32-bit access all over memory,
>Linus mentioned.

I can't backup this fully. I was testing, whether it's possible to combine
3-chip and 9-chip SIMM modules. Linux hadn't any problems (i tested it under
X with several comiling tasks running parallel and all 20MB of memory used).
OS/2 kills my machine about 20 minutes after boot with a parity fault.

Clemens Huebner
fset@guug.de