[BUG] CPU temperature not being detected

[hecto600]

the command gopsuinfo -i t not showing cpu temperature. See attachment below with lmsensors comparison:

Ps:
I'm using intel processor i7-4790k and archlinux up to date

[nwg-piotr]

I see 27 degrees in the output.

[hecto600]

Sorry for not clarifying. It shows 27, but for sure it isn't the actual cpu temperature. It doesn't change even if I open games which normally gets 60-80 degrees when used with lm_sensors sensors, kde or waybar. When the system is idling I have seen the CPU get as low as 45 degrees, but never 27. Somehow the temp showing in gopsuinfo -i t is similar with sensors category ACPI (I don't know what it is or what is capturing but it is always 27 degrees, regardless of the stress the CPU is under)

[nwg-piotr]

Well, the code returns what it finds as coretemp when acpitz not found or equal zero.

gopsuinfo/gopsuinfo.go

Line 78 in 940e8b9

if v, ok := vals["acpitz"]; ok {

. This could be inverted, of course, but I wonder if it's going to spoil the temperature indication on other processors.

[hecto600]

It is working now, thanks :). It seems though it only captures cpu core 0 temp:

If it isn't too much trouble could it be the average cpu temperature?

[nwg-piotr]

It should take the coretemp_packageid0_input first, if it exists. Wish I had the coretemp sensor in my machine... OK, I'll try to only check the coretemp_packageid0_input value in main loop, and calculate an average cores temp value outside the loop, if it fails.

[hecto600]

Can I help you testing somehow? I don't understand much about the process but I could run your script and give the results

[nwg-piotr]

Sure. I'll be able to get back to my desk later tonight. I'll let you know how to test it in a simple way.

[nwg-piotr]

OK, you can test it by building & running from the coretemp-all branch. You need to install git and go packages first.

git clone https://github.com/nwg-piotr/gopsuinfo.git -b coretemp-all
cd gopsuinfo
go run . -i t

This may take some time at first run.

[hecto600]

It seems to be working as expected now, here are some results:

[nwg-piotr]

Same here (I'm at the office, on Intel CPU). Last night I tested on an old, 2-core i3 in my ancient testing laptop. I think I could release already.