1 |
On Fri, 2013-03-22 at 20:58 +0100, Sascha Cunz wrote: |
2 |
> In the recent thread "system freezes during compiles", Carlos Henderson showed |
3 |
> the output of $(sensors), among them the output of the k10temp-pci-00c3. |
4 |
> |
5 |
> I stopped trusting that sensor. After some hours of idle, it shows me: |
6 |
> |
7 |
> k10temp-pci-00c3 |
8 |
> Adapter: PCI adapter |
9 |
> temp1: +16.8°C (high = +70.0°C) |
10 |
> (crit = +70.0°C, hyst = +67.0°C) |
11 |
> |
12 |
> 16.8°C is roughly 4°C _below_ the room temperature. If I'd boot to Windows |
13 |
> right now, the mainboard-manufacturer's system utility shows a CPU temperature |
14 |
> of 25°C (which is more or less the usual value it shows for Windows when |
15 |
> idle). |
16 |
> |
17 |
> My CPU is: |
18 |
> |
19 |
> processor : 5 |
20 |
> vendor_id : AuthenticAMD |
21 |
> cpu family : 21 |
22 |
> model : 1 |
23 |
> model name : AMD FX(tm)-6200 Six-Core Processor |
24 |
> stepping : 2 |
25 |
> microcode : 0x6000626 |
26 |
> |
27 |
> running on an Asrock mainboard (Kernel is a gentoo-3.8.1-r1). |
28 |
> |
29 |
> My question to the list is: Could this strange temperature be realistic? Does |
30 |
> anyone have similar observations with AMD-Bulldozers? |
31 |
> |
32 |
> Sascha |
33 |
> |
34 |
|
35 |
Hi Sascha, |
36 |
|
37 |
As a comparison, here's my machines processor: |
38 |
|
39 |
vendor_id : AuthenticAMD |
40 |
cpu family : 16 |
41 |
model : 10 |
42 |
model name : AMD Phenom(tm) II X6 1100T Processor |
43 |
stepping : 0 |
44 |
microcode : 0x10000bf |
45 |
|
46 |
From the kernel document hwmon/k10temp has a couple points: |
47 |
|
48 |
1. The driver detects the mainboard socket type and not the processors |
49 |
actual capabilities. If an AM3 processor is on an AM2+ mainboard, |
50 |
driver options may be required. |
51 |
|
52 |
2. The temperature reported is relative to the point at which the system |
53 |
must supply the maximum cooling for the processors specified maximum |
54 |
case temperature and maximum thermal power dissipation. |
55 |
|
56 |
It would appear that the temperature reported is quite possibly correct, |
57 |
although not intuitive or easily verified as valid. |
58 |
|
59 |
Hope that helps. |
60 |
|
61 |
Regards, |
62 |
Carlos |