[Beowulf] lm_sensors and clusters and wrong intel cpu readings

[David N. Lombard]

On Thu, Aug 09, 2012 at 12:40:23AM +0200, Vincent Diepeveen wrote:
> hi,
> 
> How do you guys monitor the CPU core temperatures?

Nowadays, I *first* look in /sys. For example, this file is on my WS:

	/sys/devices/platform/coretemp.0/temp2_input

This the temp (in thousandth's of a deg C) of /one/ of the cores.
If you look in the coretemp.0 directory (where 0 is one of the sockets,
you'll see other pseudo files, like the alarm et al.

The names may well vary depending on platform and kernel version.

> Maybe someone still has the lm_sensors script lying around somewhere  
> fixing it for the intel Xeons?

Today, I'll only use lm_sensors for things the kernel doesn't explicitly
export in /sys or /proc.

-- 
David N. Lombard, Intel, Irvine, CA
I do not speak for Intel Corporation; all comments are strictly my own.