One way to demonstrate this difference is to read the temperature directly via two methods: As root And with via a sudo prefix, where the sudo code is rather rdmsr -a -bitfield 22:16 -u sudo rdmsr -a -bitfield 22:16 -u 0x1B1Ħ0Now, the actual temperature is 100 (for my processor) - those numbers, or 35 degrees and 43 degrees. The difference is the higher amount of code executed before the sample is taken during the the logon process causing a momentary increase in processor package temperature. They are both current at the moment they are sampled.

Thank you!įor you who wanted to know what I'm talking about:ĭisagree. It's just a difference of what is current at this moment (the last recorded temp), or what was the highest temperature recorded since uptime. The files contained in the paths shown in that code in red, contain snapshot temperature statistics stored for the CPU. Yield ThermalZone(thermal_zone_path, zone_name) Thermal_zone_path = '/proc/acpi/thermal_zone'įor zone_name in sorted(os.listdir(thermal_zone_path)): To short-cut to the chase, Python script 'sysstats.py' uses this functionĭef get_thermal_zones(thermal_zone_path=None): MOTD is a series of scripts, which calls other scripts to put together the information. That is why the temp reported by the MOTD was noted as being higher.

The Temp in lm-sensors is what it is currently. The Temp reported by the MOTD when you log in via SSH, is the highest temperature logged in the stats. You just need to understand which temp is reported by what, and what they mean. Would you like the short summarized version or longer, detailed technical answer?īoth temperatures you are referring to (MOTD and lm-sensors) are correct. (When I installed lm-sensor I did modprobe nct6775 and then sensors-detect.)Īlso wonder, how does Ubuntus 'built in' fetch the temperature? What package is used? lm-sensors say my cpu is 33c and Ubuntus 'built in' says 47c. When I log in to my Ubuntu server with ssh it shows the temperature.

Ubuntu monitor cpu temperature