Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 17 Jul 1997 22:51:04 -0700
From:      David Greenman <>
To:        Howard Lew <>
Cc:, stable@FreeBSD.ORG, freebsd-questions@FreeBSD.ORG
Subject:   Re: mw fails even more... 
Message-ID:  <>
In-Reply-To: Your message of "Thu, 17 Jul 1997 19:50:58 PDT." <> 

Next in thread | Previous in thread | Raw E-Mail | Index | Archive | Help
>Can someone with a bad K6 try using more conserative settings (486) and
>disable the npx flags?  Does that change anything?

   I've always had the npx flags set to disable the Pentium optimized copy

>I just checked AMD's web site, and under the reseller corner, it says:
>AMD Recommended Thermal Solutions
>   Must use heatsink grease (in boldface)   
>I think this may tell us that the K6s are probably clocked & rated at the
>maximum speed, so that overclocking is no longer possible.  
>Funny, because I use to hear so many stories about how well a 166
>overclocks.  But I think Dave said that the cpu is not dead and that it
>goes for about 1.5 hours before it starts having trouble with random
>errors.  So my guess is a heat related problem. 
>So David, I have access to heatsink grease and can send a squirt of it
>with a new ball bearing heatsink-fan (the kind that is designed for 
>a hot Cyrix 6x86 but is a screw kind not clip-on) if you want to 
>give it a try.  I guess there is nothing to lose.  Don't make the K6 a 
>piece of jewelry yet okay? :-)
>Email me your postal address and I'll send it out (no charge). 

   Thanks for the offer, but I wouldn't consider putting together a machine
these days without thermal compound on the heatsink/chip. In other words,
the chip has always run with thermal compound. As for the heatsink/fan, it's
a screw-on type and, coupled with the thermal compound, should keep the part
cool enough. On the other hand, this is good advice in general and may help
some of the other people having problems.
   I just went in and checked the temperature of the heatsink after the
machine had been compiling for awhile (until it failed) seems to be
pretty warm, but not burn-your-fingers warm. Since it takes more than an
hour of computing before the problem surfaces, it does seem like it's a
heat related failure. It might be worth the trouble to get a Peltier effect
cooler and see if keeping the chip cold improves things any.


David Greenman
Core-team/Principal Architect, The FreeBSD Project

Want to link to this message? Use this URL: <>