Jason Gunthorpe wrote:
>
>
>On Thu, 17 Jul 1997, Kevin M. Bealer wrote:
>
>> After trying all sorts of measures, I have found one that seems so far
>> to reliably fix the problem: run it with the case off. Since there is
>> unlikely to be any "grounding problems" because of the way the case is
>> laid out (its a full tower, the case is really just a "wrapper" for a
>> solid frame, and touches no components), it is almost certainly
>> thermal.
>
>Hehe, bet the gigabyte isn't using a switching power supply :> Find the
>voltage regulators, very large heat sinks off the MB and touch them, they
>should be excessively hot (Cant keep your finger on them). You might want
>to put a fan near/on/over them to cool those puppies off. I know my MB has
>3 large heatsinks on the voltage regulators, they get hotter than the CPU!
>
>I also have a P5 MB with a switching power supply built in, it has no heat
>sinks and still carries the same load..
>
>Jason
>
The system has been up for "32 minutes" according to uptime, and they
are actually /not/ noticeably hot, just slightly above "cold metal"
temperature. Large shiny pieces of copper-colored metal, large enough
to /look/ like they should be hot. But right now, they're not.
OTOH, maybe during a kernel compile they warm up. The set6x86 thing
sets the chip to "suspend on halt" which is supposed to cool things
down a lot. Total CPU usage since bootup is 5.8 %.
So far it is "behaving itself" now that I have moved the ribbon cables
from directly in front of the fan.
[EMAIL PROTECTED]/GNU--1.3---Linux--2.0.30---
*Don't* background it, just sit there looking at it. Tell
everyone "I'll get back to you as soon as this here job is
finished". Very relaxing.
-- Lars P. Fischer
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
