Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 23 May 2007 07:09:23 -0500
From:      Eric Anderson <anderson@freebsd.org>
To:        Nate Lawson <nate@root.org>
Cc:        freebsd-acpi@freebsd.org
Subject:   Re: CPU temp AC vs Battery
Message-ID:  <46542EF3.7030809@freebsd.org>
In-Reply-To: <46379F42.3040700@freebsd.org>
References:  <4630BCC4.10601@freebsd.org>	<4637860D.8060603@freebsd.org>	<46378F75.6020007@root.org>	<4637944C.5000709@freebsd.org> <46379F42.3040700@freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 05/01/07 15:12, Eric Anderson wrote:
> On 05/01/07 14:26, Eric Anderson wrote:
>> On 05/01/07 14:05, Nate Lawson wrote:
>>> Eric Anderson wrote:
>>>> On 04/26/07 09:52, Eric Anderson wrote:
>>>>> Hi everyone,
>>>>>
>>>>> I've just noticed something very odd.  On my Dell D820 laptop, when
>>>>> running off AC power from boot, by idle CPU temperature sits around 58C.
>>>>>
>>>>> If I unplug the power, and then plug it back in again, it will drop
>>>>> down to around 49-50C within a minute or two.  It will stay there. 
>>>>> With or without powerd running.
>>>> Another note:
>>>>
>>>> If I boot up without the AC adapter plugged in, it still runs hot.  Only
>>>> the transition from AC -> battery seems to make a difference.
>>>>
>>>> Anyone with some ideas??
>>> Does the temp change at all or is it stuck at 58C?  If stuck, maybe the
>>> reading is incorrect and something in the AC line transition kicks the
>>> EC back into operation.
>> The temp does change, about 10C.
> 
> 
> Hmm.. Seems also that my performance is reduced quite a bit.  Doing some 
> rather lame CPU benchmarks (ubench -c -s), seems that I get a score of 
> around 200k on AC before unplugging, and about 104k after 
> unplugging/plugging back in.  It definitely feels slower too..
> 
> I don't see any speed changes or anything obvious in sysctl output.
> 
> 
>>> If it changes, then perhaps something is generating a lot of interrupts
>>> (perhaps SMI or SCI irqs).  More debug prints from the acpi-ca Notify
>>> routine caller would help zero in.
>>>
>> Just add some printfs in there and recompile/reboot?

Turns out that adding this to my /boot/loader.conf resolves it:

cpufreq_load="YES"


Eric






Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46542EF3.7030809>