I do not think Cputemp works with my AMD system.

It shows my temp at 34 degrees F.

I am using the kxtemp module.

My motherboard is a Gigabyte F2A68HM-H.

The most reliable temperature monitoring figure is that which is declared on the BIOS. If your temp-reading application does not match on AMD machine, just reboot to the BIOS configuration and examine the temperature values there.

Or you should find an app that matches your architecture.

Hope that helps. Good luck.

m.m.

Cputemp and lm-sensors are the only apps that work in Slako Puppy.

Checking temp by using BIOS works, but is a clunky way.

Curious...

Does anyone using any distro and has an AMD board, have a program that shows the right CPU temps ?

It's been quite a few years since I messed with lm-sensors, but IIRC it wasn't too hard to get it to show the right data. The kernel module only provided raw data from whatever sensors it found, and not necessarily in the desired units. The software that actually used/displayed the data would have a config file *somewhere* with offsets and multipliers to convert to something readable.

I did set up an AMD box once (a spanking new Athlon 2200+ ), and the last part of setting up the sensors involved noting the values as they were displayed, then quickly rebooting and comparing them to the bios values to work out the necessary factors. Clunky, yes, but the factors were mostly small integers so it wasn't too hard to work out.

i HAD an AMD mobo, and it ran debian wheezy (stable), jessie (stable) and most recently archlinux, and i never had any problems displaying cpu temps.
most recently, i used lm_sensors, with the command "sensors".

are you sure it's in fahrenheit? 34 would be a reasonable temp in celsius.

show some output.

I am running Ubuntu-Mate and it uses Psensor which uses lmsensor.

It is showing 34 degrees F.

BIOS shows cpu temp at about 120 degrees F.

I think it is using the wrong module.

I'm "fairly sure" it's not accurate. But right now my Conky shows -8C for cpu temp.
PCLinuxOS64, AMD A4-7300 APU with MSI board.

sensors-detect get me this--

Now follows a summary of the probes I have just done.
Just press ENTER to continue:

Driver `to-be-written':
* ISA bus, address 0x295
Chip `Fintek F71868A Super IO Sensors' (confidence: 9)

Driver `k10temp' (autoloaded):
* Chip `AMD Family 15h thermal sensors' (confidence: 9)

Note: there is no driver for Fintek F71868A Super IO Sensors yet.
Check http://www.lm-sensors.org/wiki/Devices for updates.

FWIW all temperature monitoring is clunky since the sources are varied and microscopic "zones'. At best, it is analogous to measuring your car's engine temperature with a sensor taped to the outside of your intake manifold whose data wire runs to a trailer housing the conversion logic circuits 100 feet behind your car. That said engineers take that into consideration and safe temp windows are reported based on that data, whether actually accurate or not. So it does work, after a fashion.

The BIOS is the least clunky since it was designed for that set of hardware, assuming decent contact of the actual sensor with the design surface, usually the underside, not the best location but as practical as it gets. Therefore th4e best solution is to use the BIOS info as a ballpark figure and assume that it is higher in some locations.

Obviously 34 F is erroneous unless you have cryo-cooling. Use the BIOS data and since all monitoring software has Offset adjustment, adjust to suit. This assumes you have determined and have the correct module setup. Check your mobo manual and verify that semsors-detect has loaded the appropriate module.

Interesting.

No big deal as I know it's running at a good temp.

Much lower than my laptops.

If you run sensors on an AMD system you get about 3 or 4 temperatures none of which are "real". I've moaned about this before on countless occasions.
install lm-sensors as mentioned above, then run sensors-detect as root and run sensors and you'll see.

Yeah, one of those things we have to live with.

If I knew where the configuration file was, I good add a correction factor.

i.e. reported temp + 90 degrees = actual temp

It's not even that -- at least one of the temperatures isn't relevant at all until it's above a certain threshold then it can, apparently, be used to tell that a system it overheating. But which of the three temperatures that it I've never found a reliable answer for.
Personally I've overclocked my processor as the temperatures don't look any higher that way and ignore the temperature. Since I've an AM3+ socket and an FX-8120 I may upgrade to an FX-8370 soon but after that I'm done with AMD. It may seem a petty reason but when their warranty says you must use the stock cooler or it's invalid and they can't give a straight answer on the maximum temperature but the one which they give is exceeded whenever the CPU does anything remotely taxing (this is with an after market cooler also) I'll not buy their products again.
Sorry for the rant, I just find it disgusting that a company doesn't provide adequate technical information.

https://wiki.archlinux.org/index.php...rature_Offsets

Here are the config files I made.
But I made some mistake.

The files are supposed to add 47 degrees C to reported value.

(I think programmers spend less time with AMD software since most folks buy Intel.)

Isn't that for Intel though? AMD sensors don't report actual temperatures but some number you're supposed to compare to some specifications they don't give.
For example, my three system temperatures are 21C, 37C and 33C (roughly 70, 99 and 91 F) but under load the first will rocket up to similar to the others but one will stay almost the same for a while then go up.