On Mon, 7 Feb 2000, John P . Looney wrote:
> On Mon, Feb 07, 2000 at 08:19:35AM -0800, Jim McEver mentioned:
> > My experience has been that the APIC errors are related to oc'ing per se
> > more that voltage. I bumped mine to [EMAIL PROTECTED], and the log was full of APIC
> > errors. I ran for awhile at 2.1v with the same results, then back down to
> > 2.0v. I did the upgrade(?) to the qq beta, and the frequency dropped
> > significantly, but there were still hundreds of errors reported over a
> > period of six hours or so last night, culminating in a hard lock. I dropped
> > back to defaults, and so far (just did this this morning, so I wouldn't call
> > it conclusive), no APIC errors - they would begin almost immediately before.
> > Backups (BRU 15.1) seem to generate a lot of them, too.
>
> You sure ? I'm getting an APIC error about every 30 mins or so. It's
> 2x400Mhz, not overclocked at all. The machine isn't very stressed - every
> ten mins or so, I'd do a big parallel make, but that wouldn't take more
> than a minute.
>
> wmbp6 reckons that the CPUs are running at 31/31.5C, case internal is
> 38C. That said, I never saw these till I stuck in a TNT2 video card.
> Wonder could they be power related...
>
This is strange... as long as i don't pump the voltage a get very few apic
errors. i.e. i can run at 550/2.0V for 24 hours getting only 2 or three
apic errors. As soon as I increase the voltage (which i only do if i clock
more, which is kinda pointless) the APIC frequency shoots up! My config:
BP6 with metal handles (rev 1?)
2x week 17 celery 366 gteed to [EMAIL PROTECTED]
2x leufkins peltiers + lots of fans.
asus TNT
realtek NIC
SBAWE32
ALL using different IRQs.
temps: idle (29/32/37) load (2 or 3 degrees hotter)
=====================================
Tim Angus [EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
ICQ UIN: 38321905
=====================================
--
=- To unsubscribe, email [EMAIL PROTECTED] with the -=
=- body of "unsubscribe linux-abit". -=
