Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 04 Nov 2003 19:55:08 -0500
From:      Jud <judmarc@fastmail.fm>
To:        nw1 <network101@covad.net>, freebsd-questions@freebsd.org
Subject:   Re: Overheating attributed to Freebsd --sysctl variables not available--
Message-ID:  <oprx43h6qr0cf2rk@mail.messagingengine.com>
In-Reply-To: <015601c3a32f$e581b820$0300a8c0@install>
References:  <015601c3a32f$e581b820$0300a8c0@install>

next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 4 Nov 2003 19:01:00 -0500, nw1 <network101@covad.net> wrote:

[snip]
>>   "Why mother board vendors release their products with such BIOS
>>   settings?
>
> What's the reason? <nothing here>

Not sure if this is what you were asking, but the reason is explained in 
the quoted lines just below:

>> Well, there is a reason: There is a possibility to get the system
>> unstable
>> and/or even to hang or crash the system."
[snip]

> What I've understood from this particular posting is this:
> the motherboards used for the AMD Athlon/Duron(s) have a default BIOS 
> setting --not
> allowing the processors to go into 'Power Saving Mode' --while intel 
> based --default BIOS
> settings: *do allow 'Power Saving Mode' for the processor(s). y/n?

I don't know enough to be able to tell you whether Intel-based mobos are 
configured by default to allow a "power saving mode," or if the fact that 
Intel CPUs run cooler (I remember reading a memorable description of the 
Athlon Thunderbird as a "blowtorch on the head of a pin") means "power 
saving mode" never becomes an issue.

> That seems to be just a bit disturbing.  If we could for the purposes of 
> this paragraph
> alone, suspend the notion of *over heating while idling*...  As someone 
> stated earlier in
> the thread, what about these cpu's under heavy load within FreeBSD? <-- 
> in our case,
> without the air-conditioner on;--> will overheat and shutdown  I have 
> put these cpu's
> under heavy load in a non FreeBSD environment and the hardware refuses 
> to break down,
> overheat or shutdown
>
> -- Unsuspend the notion of *over heating while idling*...  --
>
> Should we turn our air-conditioner off and set:
> machdep.apm_suspend_delay: 1
> machdep.apm_standby_delay: 1
> and let this dual AMD idle, the machine will overheat and shutdown in a 
> matter of hours.
> Once the room is at a tempurature warm enough to make the machine 
> shutdown, the only way
> to keep that machine on for more than five (5) to Twenty (20) minutes is 
> to turn the
> air-conditioner on and leave it on.
>
>
> Should we turn our air-conditioner off and set:
> machdep.apm_suspend_delay: 0
> machdep.apm_standby_delay: 0
>
> Run a script to loop:
> make buildworld && make buildkernel KERNCONF=<MY_KERN>
> The machine seems to run like a champ without overheating/shutting down.
>
> SUMMARY:
>
> I hear all of what you're saying here, however, if we leave apm out of 
> this altogether,
> whether inTel, AMD or any other processor, shouldn't the processor(s), 
> dual or not, be
> able to run full-throttle or, idle without overheating/ shutting down?

Yes, but: Let's say for some reason (e.g., heat conducting cement/paste 
interface not quite so nicely done on this particular unit) that this unit 
runs hot in the first place.  I don't know if the Other OS is tuned to the 
CPU, or has hooks into low-level functions in the CPU, or the CPU is tuned 
by the manufacturer to the Other OS, so that it remains just *this* side 
of shutdown with the Other OS, and goes just *that* side of shutdown with 
FreeBSD.  (We're way beyond my level of knowledge here, so if any of this 
really is correct I just got lucky.:)

Jud



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?oprx43h6qr0cf2rk>