> On Fri, 17 Dec 1999, Harold Oga wrote:
>
> > One thing I noticed on my BP6 system with dual 400's was that with a 250W
> > power supply I would get lots of APIC errors even at 66MHz and my I/O voltage
> > would fluctuate as much as 15%. Once I replaced it with a 300W power supply,
> > the voltage smoothed out, and I noticed that I rarely, if ever get APIC errors
> > at 66MHz anymore. Another thing to consider anyways.
>
> I am now hearing similar things with respect to power supply issues. Is
> it possible that the larger power supply is providing cleaner power which
> is improving the situation. While I have not checked the actual power
> usage requirements of the BP6 with it running two CPUs at 100MHz FSB, I
> have a difficult time believing that (A) the combination needs more than a
> 250W power supply in the average machine, and (B) an increase of a mere
> 50W improves the situation.
>
> Unfortunately, I am not an expert on motherboard design, nor on SMP in
> general. I do not have the facilities to test or confirm any of these
> findings. I have left that to others, but it appears they're a rare
> breed. It sure would be nice if ABit could do something to help us out
> and answer some questions.
>
> All I know is that I started to worry when I first noticed these APIC
> error messages in the log in 2.3.18 (or thereabouts).
(When the APIC code was modified to log errors.)
It might be worth noting that I'm running two SLOT 1 Celeron 300As
(drilled and soldered to work) on an ASUS P2B-DS board, and I can
overclock them both to 463 MHz (103 MHz FSB) with no problems. I have
never seen an APIC error message ever, and I've been running these at
300, 450, and 463 quite regularly (while compiling X and other things, as
well). I'm also using a fairly cheap power supply, and I can see the
voltage drop quite easily when the CPUs are not idling.
Simon-
[ Stormix Technologies Inc. ][ NetNation Communcations Inc. ]
[ sim@stormix.com ][ sim@netnation.com ]
[ Opinions expressed are not necessarily those of my employers. ]
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/