Hello,

Sven Anders wrote:
> Nicolas Boichat schrieb:
>> Sheer El-Showk wrote:
>>> Yes, this is quite reminicent of my experience with the fans.  I
>>> thought it might be lag on stuff that I did earlier but I'm not sure.
>>> Nicolas has suggested that the SMC uses additional temperature
>>> monitors that are not exposed by the applesmc module.
>> Not what I said (or maybe not what I meant). I said the fan are very
>> probably controlled by temperature sensors connected to the SMC, but not
>> the sensors inside the CPU (i.e. not the values reported by coretemp).
> 
> If you read the values from /sys and compare them with the 'coretemp' values,
> the value from sys is always a little bit higher.
> So I think your are right here.
> 
> FYI: From a sensor utility under OSX I had the following output:
> 
>  CPU A: 55°C               (temperature_1)
>  Heatsink A: 47°C          (temperature_2)
>  GPU Heatsink: 51°C        (temperature_3)
>  Heatsink B: 39°C          (temperature_4)
>  GPU: 45°C                 (temperature_5)
> 
> (I tried to match the values with the applesmc temperatures, can you confirm
> these?)

Which utility are you using on OSX?

> What is the temperature_0, any idea? Reading temperature_6 gives an error.
> What could the 'Heatsink B' mean? I only have one CPU!
> My harddisk temperature is about 36°C.
> 
> If the temperature of the GPU reaches 45°C, then fans begin to speed up.
> 
> I think the GPU is one reason for this behaviour, but it seems the CPU
> runs a litte bit cooler under MacOSX. Maybe, because it's using the
> advanced power-saving states of the CPU?

That's what I think.

> Maybe a later kernel (with support for the advanced states + dynticks +
> auto-usb-suspend) will make this better...
> 
> 
>> I did some debugging of the OS X kernel a while ago, and the only
>> sensors OS X is polling is the light sensors, to adjust the keyboard
>> backlight.
> 
> I read (some time ago), that the SMC might trigger an interrupt and that
> these could prevent the driver from polling the values. Is this true?
> In the same artice the person the applesmc responsible for power-drain,
> which I could not believe...

Where? Who wrote that? Interesting, if the person who wrote the article
is well-informed...

> Is the applesmc driver constantly reading the values or does it read it
> only on demand (i.e. by reading /sys for instance)?

On demand, except for the input device that can be read constantly
(possibly only when the device is open, I'm not sure, it's a part of the
code that I copy-pasted from hdaps (easy to check if you want, just add
some printk at the right places)).

> I'm asking this, because I'm getting these messages randomly:
>   applesmc: wait status failed: 5 != 0
> and I'm not sure if a process using the module.

I get those too. They are harmless.

Best regards,

Nicolas

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
Mactel-linux-users mailing list
Mactel-linux-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/mactel-linux-users

Reply via email to