dmesg showing wrong frequency (IBM T30)

[Kevin Oberman]

> Date: Fri, 18 Jul 2003 19:24:20 +0200
> From: Tobias Roth <roth at iam.unibe.ch>
> Sender: owner-freebsd-mobile at freebsd.org
> 
> On Fri, Jul 18, 2003 at 06:22:27PM +0200, Poul-Henning Kamp wrote:
> > What's "wrong" here is that the BIOS/ACPI firmware in your laptop
> > runs your CPU at a reduced rate in order to make the battery last
> > longer.
> 
> it should NOT do this. I set the bios to disable speedstep and to
> 'max performance' while on AC. also, i run apm and not acpi.
> 
> > It can also be that the case that the "cooling solution" (ie: fans,
> > fins etc) does not work well enough and the ACPI code has slowed
> > down the CPU in order to not melt anything [*].
> 
> that was also one of my suspicions. the fan is brand new, ibm replaced
> it after i sent the unit in because of heat problems.
> 
> could it be that the design of this laptop is bad, when it comes to
> heat conduction?
> 
> shouldn't i get the full 1.8GHz when running on AC, and be able to
> do a buildworld without any sig11 on a hot summer day (~28 deg c room
> temp). those sig11 are just another thing that may be connected to my
> low cpu freq...

I have not had any issue running a buildworld on my T30 on a hot
day. (It hit 105 at my house yesterday and 103 at Lawrence Livermore
about 3 miles away.)

It runs at 1.8 GHz and takes under an hour to build CURRENT,
about 35% longer than it took with STABLE. No 11s sighted. It does get
HOT..too hot to really be a laptop unless you have well insulated
pants, but within spec for temperature. (At least with ACPI running
so that I can monitor the temperature.) The exhaust vent is blowing a
strong stream of very warm air.

I do see messages that the system is switching between performance and
economy mode when I switch from AC to battery, but the CPU pays no
attention and the speed stays unchanged under both APM and ACPI. :-(

It sounds like something is very wrong with your T30.
-- 
R. Kevin Oberman, Network Engineer
Energy Sciences Network (ESnet)
Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
E-mail: oberman at es.net			Phone: +1 510 486-8634