You are not logged in.

#1 2013-04-07 09:52:32

berbae
Member
From: France
Registered: 2007-02-12
Posts: 1,142

Erroneous sensors outputs

$ sensors
nouveau-pci-0080
Adapter: PCI adapter
temp1:       +6552.0°C  (high = +95.0°C, crit = +145.0°C)

coretemp-isa-0000
Adapter: ISA adapter
Core 0:       +58.0°C  (high = +76.0°C, crit = +100.0°C)
Core 1:       +37.0°C  (high = +76.0°C, crit = +100.0°C)

temp1 is obviously wrong, but varies in time; what is it for?.
Core 0 doesn't vary, it's always 58°C.
Core 1 varies, but what is its meaning and is it correct?

$ lsmod|grep -e coretemp -e nouveau
coretemp                6198  0 
nouveau               925878  2 
button                  4701  1 nouveau
video                  11363  1 nouveau
mxm_wmi                 1467  1 nouveau
wmi                     8379  2 mxm_wmi,nouveau
i2c_algo_bit            5391  1 nouveau
drm_kms_helper         35517  1 nouveau
ttm                    64611  1 nouveau
drm                   226103  4 ttm,drm_kms_helper,nouveau
i2c_core               23171  5 drm,drm_kms_helper,i2c_algo_bit,nouveau,i2c_nforce2

My motherboard is 'Gigabyte Technology Co., Ltd. GA-73PVM-S2H' with 'CPU0: Intel Pentium(R) Dual-Core  CPU      E5200'.
I don't see how to configure that and get exact values to monitor hardward heat with nouveau driver.
Help would be appreciated.

Offline

#2 2013-04-08 01:15:52

felipense
Member
Registered: 2013-04-05
Posts: 62

Re: Erroneous sensors outputs

It's a noisy cpu? What temperatures you have in Windows?

Offline

#3 2013-04-08 14:53:32

berbae
Member
From: France
Registered: 2007-02-12
Posts: 1,142

Re: Erroneous sensors outputs

I take the 'Core 1' value from coretemp isa to be the CPU temp in Conky.
I take the 'temp1' value from nouveau pci to be the GPU temp; but I apply an offset to it.
I had to make a guess for the offset value:
when the computer boots up, the value is 6616°C, so I take that to be 40°C (just a little higher than the CPU).
And in .conkyrc I use: ${hwmon 0 temp 1 1.0 -6576.0} for the GPU temp.
But it's just a guess; I could not find more precise infos on the web about how to get more reliable values.

With that, the GPU temp is shown to constantly change, when idle, in a 10°C range (between 40 and 50°C).
Does this seem normal? Does someone, monitoring the GPU temp, see that also?

I hope that the above choices will at least give me an idea about temp variations...but I am not sure how reliably though.

I still would appreciate any other suggestions. Thanks.

Offline

#4 2013-04-08 17:20:05

x33a
Forum Moderator
Registered: 2009-08-15
Posts: 3,424
Website

Re: Erroneous sensors outputs

You can give nvidia-smi a try, but AFAIK it only works with the binary driver.

Offline

#5 2013-04-09 10:20:08

berbae
Member
From: France
Registered: 2007-02-12
Posts: 1,142

Re: Erroneous sensors outputs

I see that I cannot use reliably the temp1 value from nouveau pci, because the offset seems not to be the same with each reboot.
Today I try '${hwmon 0 temp 1 0.01 0.0}' to take the two digits to the left; but now the value doesn't change, even when running a streaming video; and I don't think it is correct.
To use it, I would have to know the formula to translate it to the right GPU temp value, if it is even possible...
It's sad, because the GPU temperature is certainly the most important one to know with the nouveau driver, to detect overheating.

For now I will see if I get power-off from overheating, even when using an additional external cooling fan; and if it is the case, I will have to switch again to the proprietary nvidia driver.

Edit:
In fact the temp1 value seems only to need some time to stabilize at the value I take to be 40°C GPU idle temperature;
so I will use again '${hwmon 0 temp 1 1.0 -6576.0}' in .conkyrc because the variations then seem to correspond well to the GPU usage, and consequently that gives me an approximation of the GPU heat.

Edit2:
I changed the GPU (GeForce 7100) idle temperature to a seemingly more realistic value of 50°C,
with '${hwmon 0 temp 1 1.0 -6566.0}'.

Last edited by berbae (2013-04-10 14:28:02)

Offline

Board footer

Powered by FluxBB