[lm-sensors] Sensors [via-cputemp] is cpu intensive?

lmsensors at continuingtime.com lmsensors at continuingtime.com
Sun Apr 22 19:03:03 CEST 2012

On 4/22/2012 2:42 PM, Jean Delvare wrote:
> Hi Dave,
> On Sun, 22 Apr 2012 11:19:24 +0100, lmsensors at continuingtime.com wrote:
> Most of /proc/acpi has been deprecated over time in favor of equivalent
> (but often more generic) sysfs attributes. For
> example /proc/acpi/thermal_zone/THRM/temperature would now
> be /sys/class/thermal/thermal_zone0/temp. If your kernel had
> CONFIG_THERMAL_HWMON enabled, you would also see this thermal zone in
> the output of "sensors".

I don't have anything of note in /sys/class/thermal so I'll look at what 
I might be missing from the newer acpi configs.

> I think you're looking at the numbers in the wrong way. It's not 
> sensors or grep or anything causing a 10°C increase in temperature. 
> It's the CPU being nicely designed with efficient low-power C-states 
> which make it possible to save 10°C in idle state. 

Yes and no.  Even when the CPU was 'idle' and cooling down from the max 
of 57 to 46-ish, I was able to reliably increase the CPU temps by 3+ 
degC by running sensors twice and by 8-10 degC by running sensors 200 
times which, in a loop.  And running it 200 times on a 1.8GHz cpu takes 
3 seconds.  Nothing takes 3 seconds.

Yes, I realize that time is required to setup the execution environment 
for sensors and it relies on the kernel and bus wait times, etc etc 
etc...  But...

Nothing takes 3 seconds.  :)


More information about the lm-sensors mailing list