lm_sensors
lm_sensors[dead link 2024-10-12 ⓘ] (Linux monitoring sensors) is a free and open-source application that provides tools and drivers for monitoring temperatures, voltage, and fans. This document explains how to install, configure, and use lm_sensors.
Installation
Install the lm_sensors package.
Configuration
Use sensors-detect as root to detect and generate a list of kernel modules:
Enter), unless you know exactly what you are doing. See #Laptop screen issues after running sensors-detect.# sensors-detect
It will ask to probe for various hardware. The "safe" answers are the defaults, so just hitting Enter to all the questions will generally not cause any problems. This will create the /etc/conf.d/lm_sensors configuration file which is used by lm_sensors.service to automatically load kernel modules on boot.
--auto flag may be used to automatically accept all safe answers for sensors-detect.When the detection is finished, a summary of the probes is presented.
Example:
# sensors-detect
This program will help you determine which kernel modules you need
to load to use lm_sensors most effectively. It is generally safe
and recommended to accept the default answers to all questions,
unless you know what you're doing.
Some south bridges, CPUs or memory controllers contain embedded sensors.
Do you want to scan for them? This is totally safe. (YES/no):
Module cpuid loaded successfully.
Silicon Integrated Systems SIS5595... No
VIA VT82C686 Integrated Sensors... No
VIA VT8231 Integrated Sensors... No
AMD K8 thermal sensors... No
AMD Family 10h thermal sensors... No
...
Now follows a summary of the probes I have just done.
Just press ENTER to continue:
Driver `coretemp':
* Chip `Intel digital thermal sensor' (confidence: 9)
Driver `lm90':
* Bus `SMBus nForce2 adapter at 4d00'
Busdriver `i2c_nforce2', I2C address 0x4c
Chip `Winbond W83L771AWG/ASG' (confidence: 6)
Do you want to overwrite /etc/conf.d/lm_sensors? (YES/no):
ln -s '/usr/lib/systemd/system/lm_sensors.service' '/etc/systemd/system/multi-user.target.wants/lm_sensors.service'
Unloading i2c-dev... OK
Unloading cpuid... OK
/etc/conf.d/lm_sensors. Answering YES also automatically starts the service.Running sensors
Example running sensors:
$ sensors
coretemp-isa-0000
Adapter: ISA adapter
Core 0: +35.0°C (crit = +105.0°C)
Core 1: +32.0°C (crit = +105.0°C)
w83l771-i2c-0-4c
Adapter: SMBus nForce2 adapter at 4d00
temp1: +28.0°C (low = -40.0°C, high = +70.0°C)
(crit = +85.0°C, hyst = +75.0°C)
temp2: +37.4°C (low = -40.0°C, high = +70.0°C)
(crit = +110.0°C, hyst = +100.0°C)
Adding DIMM temperature sensors
To find the temperature sensors of DIMMs, install the i2c-tools package. Once installed, load the i2c-dev kernel module.
# modprobe i2c_dev
To show all the columns, use i2cdetect as root:
# i2cdetect -l
i2c-2 smbus SMBus PIIX4 adapter port 2 at 0b00 SMBus adapter i2c-2 smbus SMBus PIIX4 adapter port 1 at 0b20 SMBus adapter i2c-0 smbus SMBus PIIX4 adapter port 0 at 0b00 SMBus adapter
Otherwise, its output will appear as follows:
i2c-2 unknown SMBus PIIX4 adapter port 2 at 0b00 N/A i2c-2 unknown SMBus PIIX4 adapter port 1 at 0b20 N/A i2c-0 unknown SMBus PIIX4 adapter port 0 at 0b00 N/A
In the following example, RAM sticks are connected to the bus SMBus 0. The i2cdetect command will show the devices that are connected to the bus. The -y 0 argument uses the i2c-0 smbus. Check other buses if needed.
# i2cdetect -y 0
0 1 2 3 4 5 6 7 8 9 a b c d e f 00: -- -- -- -- 0c -- -- -- 10: 10 -- -- -- -- -- -- -- 18 19 -- -- -- -- -- -- 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 30: -- -- -- -- -- -- 36 -- -- -- -- -- -- -- -- -- 40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 4f 50: 50 51 -- -- -- -- -- -- -- -- -- -- -- -- -- -- 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 70: -- -- -- -- -- -- -- 77
RAM SPD's (serial presence detect) start from address 0x50 and RAM temperature sensors start from 0x18 at same bus. In this example, there are 2 DIMMs available. The address 0x18 and 0x19 are the DIMMs temperature sensors.
To read the temperatures of RAM sticks, we need the jc42 kernel module loaded. You need to tell the module which addresses to use. This process consists of writing the module_name and address to smbus_path. For example:
# modprobe jc42 # echo jc42 0x18 > /sys/bus/i2c/devices/i2c-0/new_device # echo jc42 0x19 > /sys/bus/i2c/devices/i2c-0/new_device
After that your ram sticks temperature will be visible:
$ sensors
jc42-i2c-0-19
Adapter: SMBus PIIX4 adapter port 0 at 0b00
temp1: +50.7°C (low = +0.0°C) ALARM (HIGH, CRIT)
(high = +0.0°C, hyst = +0.0°C)
(crit = +0.0°C, hyst = +0.0°C)
jc42-i2c-0-18
Adapter: SMBus PIIX4 adapter port 0 at 0b00
temp1: +51.8°C (low = +0.0°C) ALARM (HIGH, CRIT)
(high = +0.0°C, hyst = +0.0°C)
(crit = +0.0°C, hyst = +0.0°C)
Reading SPD values from memory modules (optional)
To read the SPD timing values from memory modules, install the i2c-tools package. Once installed, load the at24 or eeprom (deprecated) kernel module.
# modprobe at24
Finally, view memory information with decode-dimms.
Here is partial output from one machine:
# decode-dimms
Memory Serial Presence Detect Decoder By Philip Edelbrock, Christian Zuckschwerdt, Burkart Lingner, Jean Delvare, Trent Piepho and others Decoding EEPROM: /sys/bus/i2c/drivers/eeprom/0-0050 Guessing DIMM is in bank 1 ---=== SPD EEPROM Information ===--- EEPROM CRC of bytes 0-116 OK (0x583F) # of bytes written to SDRAM EEPROM 176 Total number of bytes in EEPROM 512 Fundamental Memory type DDR3 SDRAM Module Type UDIMM ---=== Memory Characteristics ===--- Fine time base 2.500 ps Medium time base 0.125 ns Maximum module speed 1066MHz (PC3-8533) Size 2048 MB Banks x Rows x Columns x Bits 8 x 14 x 10 x 64 Ranks 2 SDRAM Device Width 8 bits tCL-tRCD-tRP-tRAS 7-7-7-33 Supported CAS Latencies (tCL) 8T, 7T, 6T, 5T ---=== Timing Parameters ===--- Minimum Write Recovery time (tWR) 15.000 ns Minimum Row Active to Row Active Delay (tRRD) 7.500 ns Minimum Active to Auto-Refresh Delay (tRC) 49.500 ns Minimum Recovery Delay (tRFC) 110.000 ns Minimum Write to Read CMD Delay (tWTR) 7.500 ns Minimum Read to Pre-charge CMD Delay (tRTP) 7.500 ns Minimum Four Activate Window Delay (tFAW) 30.000 ns ---=== Optional Features ===--- Operable voltages 1.5V RZQ/6 supported? Yes RZQ/7 supported? Yes DLL-Off Mode supported? No Operating temperature range 0-85C Refresh Rate in extended temp range 1X Auto Self-Refresh? Yes On-Die Thermal Sensor readout? No Partial Array Self-Refresh? No Thermal Sensor Accuracy Not implemented SDRAM Device Type Standard Monolithic ---=== Physical Characteristics ===--- Module Height (mm) 15 Module Thickness (mm) 1 front, 1 back Module Width (mm) 133.5 Module Reference Card B ---=== Manufacturer Data ===--- Module Manufacturer Invalid Manufacturing Location Code 0x02 Part Number OCZ3G1600LV2G ...
Using sensor data
Graphical front-ends
There are a variety of front-ends for sensors data.
- psensor — GTK application for monitoring hardware sensors, including temperatures and fan speeds. Monitors motherboard and CPU (using lm-sensors), Nvidia GPUs (using XNVCtrl), and harddisks (using hddtemp or libatasmart).
- xsensors — X11 interface to lm_sensors.
- Netdata — Web based system monitor (netdata).
- CoolerControl — A program to monitor and control your cooling devices.
For specific Desktop environments:
- Freon (GNOME Shell extension) — Extension for displaying CPU temperature, disk temperature, video card temperature , voltage and fan RPM in GNOME Shell.
- GNOME Sensors Applet — Applet for the GNOME Panel to display readings from hardware sensors, including CPU temperature, fan speeds and voltage readings.
- lm-sensors (LXPanel plugin) — Monitor temperature/voltages/fan speeds in LXDE through lm-sensors.
- MATE Sensors Applet — Display readings from hardware sensors in your MATE panel.
- Sensors (Xfce4 panel plugin) — Hardware sensors plugin for the Xfce panel.
- Thermal Monitor (Plasma applet) — A KDE Plasmoid for displaying system temperatures.
sensord
There is an optional daemon called sensord (included with the lm_sensors package) which can log data to a round robin database (rrd) and later visualize graphically. See the sensord(8) man page for details.
Tips and tricks
Adjusting values
In some cases, the data displayed might be incorrect or users may wish to rename the output. Use cases include:
- Incorrect temperature values due to a wrong offset (i.e. temps are reported 20 °C higher than actual).
- Users wish to rename the output of some sensors.
- The cores might be displayed in an incorrect order.
All of the above (and more) can be adjusted by overriding the package provided settings in /etc/sensors3.conf by creating /etc/sensors.d/foo wherein any number of tweaks will override the default values. It is recommended to rename 'foo' to the motherboard brand and model but this naming nomenclature is optional.
Custom configuration files for a number of motherboards can be found in the configs directory of the lm-sensors package and be used as templates.
/etc/sensors3.conf directly since package updates will overwrite any changes thus losing them.Example 1. Adjusting temperature offsets
This is a real example on a Zotac ION-ITX-A-U motherboard. The coretemp values are off by 20 °C (too high) and are adjusted down to Intel specs.
$ sensors
coretemp-isa-0000 Adapter: ISA adapter Core 0: +57.0°C (crit = +125.0°C) Core 1: +55.0°C (crit = +125.0°C) ...
Run sensors with the -u switch to see what options are available for each physical chip (raw mode). If some of the raw labels you are presented seem not to be configurable, look at the /sys/class/hwmon directory tree. Each device mentioned there has a name file, which can be used to match the device it is referring to. And then try the labels referred to by that directory.
$ sensors -u
coretemp-isa-0000 Adapter: ISA adapter Core 0: temp2_input: 57.000 temp2_crit: 125.000 temp2_crit_alarm: 0.000 Core 1: temp3_input: 55.000 temp3_crit: 125.000 temp3_crit_alarm: 0.000 ...
Create the following file overriding the default values:
/etc/sensors.d/Zotac-IONITX-A-U
chip "coretemp-isa-0000" label temp2 "Core 0" compute temp2 @-20,@-20 label temp3 "Core 1" compute temp3 @-20,@-20
Now invoking sensors shows the adjust values:
$ sensors
coretemp-isa-0000 Adapter: ISA adapter Core 0: +37.0°C (crit = +105.0°C) Core 1: +35.0°C (crit = +105.0°C) ...
Example 2. Renaming labels
This is a real example on an Asus A7M266. The user wishes more verbose names for the temperature labels temp1 and temp2:
$ sensors
as99127f-i2c-0-2d Adapter: SMBus Via Pro adapter at e800 ... temp1: +35.0°C (high = +0.0°C, hyst = -128.0°C) temp2: +47.5°C (high = +100.0°C, hyst = +75.0°C) ...
Create the following file to override the default values:
/etc/sensors.d/Asus_A7M266
chip "as99127f-*" label temp1 "Mobo Temp" label temp2 "CPU0 Temp"
Now invoking sensors shows the adjust values:
$ sensors
as99127f-i2c-0-2d Adapter: SMBus Via Pro adapter at e800 ... Mobo Temp: +35.0°C (high = +0.0°C, hyst = -128.0°C) CPU0 Temp: +47.5°C (high = +100.0°C, hyst = +75.0°C) ...
Example 3. Renumbering cores for multi-CPU systems
This is a real example on an HP Z600 workstation with dual Xeons. The actual numbering of physical cores is incorrect: numbered 0, 1, 9, 10 which is repeated into the second CPU. Most users expect the core temperatures to report out in sequential order, i.e. 0,1,2,3,4,5,6,7.
$ sensors
coretemp-isa-0000 Adapter: ISA adapter Core 0: +65.0°C (high = +85.0°C, crit = +95.0°C) Core 1: +65.0°C (high = +85.0°C, crit = +95.0°C) Core 9: +66.0°C (high = +85.0°C, crit = +95.0°C) Core 10: +66.0°C (high = +85.0°C, crit = +95.0°C) coretemp-isa-0004 Adapter: ISA adapter Core 0: +54.0°C (high = +85.0°C, crit = +95.0°C) Core 1: +56.0°C (high = +85.0°C, crit = +95.0°C) Core 9: +60.0°C (high = +85.0°C, crit = +95.0°C) Core 10: +61.0°C (high = +85.0°C, crit = +95.0°C) ...
Again, run sensors with the -u switch to see what options are available for each physical chip:
$ sensors -u coretemp-isa-0000
coretemp-isa-0000 Adapter: ISA adapter Core 0: temp2_input: 61.000 temp2_max: 85.000 temp2_crit: 95.000 temp2_crit_alarm: 0.000 Core 1: temp3_input: 61.000 temp3_max: 85.000 temp3_crit: 95.000 temp3_crit_alarm: 0.000 Core 9: temp11_input: 62.000 temp11_max: 85.000 temp11_crit: 95.000 Core 10: temp12_input: 63.000 temp12_max: 85.000 temp12_crit: 95.000
$ sensors -u coretemp-isa-0004
coretemp-isa-0004 Adapter: ISA adapter Core 0: temp2_input: 53.000 temp2_max: 85.000 temp2_crit: 95.000 temp2_crit_alarm: 0.000 Core 1: temp3_input: 54.000 temp3_max: 85.000 temp3_crit: 95.000 temp3_crit_alarm: 0.000 Core 9: temp11_input: 59.000 temp11_max: 85.000 temp11_crit: 95.000 Core 10: temp12_input: 59.000 temp12_max: 85.000 temp12_crit: 95.000 ...
Create the following file overriding the default values:
/etc/sensors.d/HP_Z600
chip "coretemp-isa-0000" label temp2 "Core 0" label temp3 "Core 1" label temp11 "Core 2" label temp12 "Core 3" chip "coretemp-isa-0004" label temp2 "Core 4" label temp3 "Core 5" label temp11 "Core 6" label temp12 "Core 7"
Now invoking sensors shows the adjust values:
$ sensors
coretemp-isa-0000 Adapter: ISA adapter Core0: +64.0°C (high = +85.0°C, crit = +95.0°C) Core1: +63.0°C (high = +85.0°C, crit = +95.0°C) Core2: +65.0°C (high = +85.0°C, crit = +95.0°C) Core3: +66.0°C (high = +85.0°C, crit = +95.0°C) coretemp-isa-0004 Adapter: ISA adapter Core4: +53.0°C (high = +85.0°C, crit = +95.0°C) Core5: +54.0°C (high = +85.0°C, crit = +95.0°C) Core6: +59.0°C (high = +85.0°C, crit = +95.0°C) Core7: +60.0°C (high = +85.0°C, crit = +95.0°C) ...
Automatic lm_sensors deployment
Users wishing to deploy lm_sensors on multiple machines can use the following to accept the defaults to all questions:
# sensors-detect --auto
S.M.A.R.T. drive temperature
Since kernel 5.6[dead link 2024-10-12 ⓘ] the drivetemp module will report SATA/SAS temperature through hwmon, but sensors-detect does not automatically detect this so the module must be manually loaded.
# modprobe drivetemp
You should now see entries similar to this in your sensors output:
sensors
drivetemp-scsi-1-0
Adapter: SCSI adapter
temp1: +33.0°C
drivetemp-scsi-2-0
Adapter: SCSI adapter
temp1: +32.0°C (low = +0.0°C, high = +70.0°C)
(crit low = +0.0°C, crit = +70.0°C)
(lowest = +29.0°C, highest = +41.0°C)
You can now load the module at boot. Alternatively, manually add it to the HWMON_MODULES line of /etc/conf.d/lm-sensors. Do note it will not be added automatically when sensors-detect will be allowed to write this file again.
Persistent device names
Many pieces of software expect sensor devices to stay put in /sys/class/hwmon/hwmonX, but more often than not, they don't on systems with more than 1-2 devices providing a hwmon interface. Software should probably parse the hwmon?/name or use lmsensors libraries, but more often than not, they sadly don't. Some software (example: Monitorix or certain modules of it, namely amdgpu) expect persistent names somewhere else.
Hence, the following kind of udev rules might be useful. Not all software can use them (for example, KDE system monitor - which sadly, makes these software almost useless on many systems). For many cases, simply matching the hwmon subsystem and a suitable name in the udev rule should be enough - but not always! See Udev page for more information on writing rules.
One can not rename or symlink under /sys hieraerchy. A SYMLINK+= -statement will also not work. Hence, we need to use the RUN+= statement (note, the symlink does not need to be under /dev as in this example - there is no standard nor a good place for them).
/etc/udev/rules.d/99-persistent-hwmon-names.rules
# my motherboard sensor chip:
ACTION=="add", SUBSYSTEM=="hwmon", ATTRS{name}=="nct6687", RUN+="/bin/sh -c 'ln -s /sys$devpath /dev/nct6678'"
# an USB device providing sensors:
ACTION=="add", SUBSYSTEM=="hwmon", ATTRS{name}=="corsaircpro", RUN+="/bin/sh -c 'ln -s /sys$devpath /dev/corsaircpro'"
# my GPU:
ACTION=="add", SUBSYSTEM=="hwmon", ATTRS{vendor}=="0x1002", ATTRS{device}=="0x73bf", RUN+="/bin/sh -c 'ln -s /sys$devpath /dev/rx6900xt'"
Troubleshooting
K10Temp module
Some K10 processors have issues with their temperature sensor. See the k10temp documentation for more information.
On affected machines the module will report "unreliable CPU thermal sensor; monitoring disabled". To force monitoring anyway, you can run the following:
# rmmod k10temp # modprobe k10temp force=1
Confirm that the sensor is in fact valid and reliable. If it is, can edit /etc/modprobe.d/k10temp.conf and add:
options k10temp force=1
This will allow the module to load at boot.
Asus B450M-A/A320M-K/A320M-K-BR motherboards
These motherboards use a IT8655E chip, which is not supported by the it87 kernel driver, as of Nov 2020 [1]. However, it is supported by the upstream version of the kernel driver [2]. The DKMS variant is contained in it87-dkms-gitAUR.
Asus B450/X399/X470 motherboards with AM4 Socket
Some recent Asus motherboards use a ITE IT8665E chip, accessing the temperature, fan and voltage sensors may require the asus-wmi-sensors module. It is part of the mainline kernel since 5.17: load the asus-wmi-sensors kernel module which uses the UEFI interface and may require a BIOS update on some boards [3].
Alternatively, the it87 module reads the values from the chip directly, install it87-dkms-gitAUR and load the it87 kernel module.
ASUS H97/Z97/Z170/Z370i/X570/B550/B650-PLUS/X670 motherboards
With some recent ASUS motherboards, fan and voltage sensor access may require the nct6775 kernel module to be loaded.
You may also need to add the following kernel parameter:
acpi_enforce_resources=lax
See https://bugzilla.kernel.org/show_bug.cgi?id=204807 for more information.
Asrock Deskmini H470
The STX board of the Deskmini H470 uses a NCT6683 chip, for accessing the temperature, fan and voltage sensors the loading of nct6683 module is required.
For proper values of the nct6683 module have a module config file created:
/etc/modprobe.d/nct6683.conf
options nct6683 force=1
Gigabyte B250/Z370/B450M/B560M/B660M/Z690/B550 motherboards
Some Gigabyte motherboards use the ITE IT8686E, ITE8689 (for B560 and B660M) or ITE8689E (for Z690 and B550) chip, which is not supported by the it87 kernel driver, as of May 2019 [4]. However, it is supported by the upstream version of the kernel driver [5]. The DKMS variant is contained in it87-dkms-gitAUR. As with #ASUS H97/Z97/Z170/Z370i/X570/B550/B650-PLUS/X670 motherboards, a kernel parameter is required before attempting to install the module:
acpi_enforce_resources=lax
Furthermore, supply the id of the chip when loading the module as follows:
# modprobe it87 force_id=0x8686 or # modprobe it87 force_id=0x8689 # for B560 # modprobe it87 force_id=0x8628 # for Z690 and B550
Or you can load the module during boot process by creating the following two files:
/etc/modules-load.d/it87.conf
it87
For Z690 and B550
/etc/modprobe.d/it87.conf
options it87 force_id=0x8628
For others
/etc/modprobe.d/it87.conf
options it87 ignore_resource_conflict=1
Once the module is loaded you can use the sensors tool to probe the chip. Now you can also use fancontrol to control the speed step of your case fan.
Optionally installation of zenpower3-dkmsAUR may allow greater fine tuning of the motherboard's cooling system. However, it does disable the default k10temp module.
Gigabyte GA-J1900N-D3V
This motherboard uses the ITE IT8620E chip (useful also to read voltages, mainboard temp, fan speed). As of October 2014, lm_sensors has no driver support for chip ITE IT8620E [6][dead link 2024-10-12 ⓘ] [7]. lm_sensors developers had a report that the chip is somewhat compatible with the IT8728F for the hardware monitoring part. However, as of August 2016, [8] lists the IT8620E as supported.
You can load the module at runtime with modprobe:
$ modprobe it87 force_id=0x8728
Or you can load the modules during boot process by creating the following two files:
/etc/modules-load.d/it87.conf
it87
/etc/modprobe.d/it87.conf
options it87 force_id=0x8603
Once the module is loaded you can use the sensors tool to probe the chip.
Now you can also use fancontrol to control the speedsteps of your case fan.
MAG B650 TOMAHAWK WIFI (MS-7D75)
This motherboard uses a Nuvoton NCT6687-R chip for fan, voltage and temperature readings. You need kernel module nct6687d-dkms-gitAUR to access those values.
Once installed you can load the module
$ modprobe nct6687
And then check the output of sensors
$ sensors
Asrock B650M Pro RS
This motherboard uses the Nuvoton NCT6796D-S for fan, voltage and temperature readings. This exact variant of the chip has a different ID and therefore is not detected without a module config file:
/etc/modprobe.d/nct6775.conf
options nct6775 force_id=0xd801
Laptop screen issues after running sensors-detect
This is caused by lm-sensors messing with the Vcom values of the screen while probing for sensors. It has been discussed and solved at the forums already: https://bbs.archlinux.org/viewtopic.php?id=193048. However, make sure to read through the thread carefully before running any of the suggested commands.