Best of the Newsgroups: thermal problem p4
What is this stuff?
Main Index
From: Bela Lubkin <filbo@armory.com>
Subject: Re: Memory overlap
Date: 11 Nov 2005 03:37:02 -0500
Message-ID: <200511110036.aa11910@deepthought.armory.com>
References: <kNDcf.14$s14.5@newsread2.news.pas.earthlink.net> <afScf.280$s14.227@newsread2.news.pas.earthlink.net> <IprzGJ.2yw@stevedunn.ca>
Stephen M. Dunn wrote:
> In article <afScf.280$s14.227@newsread2.news.pas.earthlink.net> "SDS" <sds10@earthlink.net> writes:
> $Terrible lapse on my part. The client had a new P4 machine and was
> $installing 5.0.5 which I guess is incompatible. I never even checked.
>
> Officially, it may be, but I'm sure I'm not the only person who's
> had 5.0.5 running on a P4, and a couple of years ago, after I'd
> mentioned having chosen a PIII system due to 5.0.5 not supporting
> the P4, a helpful expert who's well known in this group emailed
> me and said that in reality there shouldn't be problems running
> 5.0.5 on a P4. (That PIII motherboard or CPU died - never figured
> out which one was responsible - and I replaced it with a P4, and
> continued to run 5.0.5 for a while before upgrading to 5.0.7.)
Very diplomatic. ;-}
Anyway, I think what I really said was that the chance of problems is
very small, but it's still an at-your-own-risk kind of situation.
There are probably ways that a runaway _application_ could cause a P4
CPU to overheat under OSR505. In fact it isn't clear to me that the
mods done between 506 and 507, intended to handle all the P4 thermal
problems, ever really did address runaway applications. If the P4
thermal design is really as I understood from Intel documents, I'm not
even sure it's _possible_ for an OS to protect against some situations.
The software changes demanded for thermal control were (in my opinion)
by far the worst feature of the P4 architecture family. Worse even than
the 30-40% loss of horsepower-per-MHz compared to the P3 architecture.
>Bela<
Get APLawrence.com by RSS
Add your comments