On the live ISO the number is 300 both in nvtop and lact, on the install immediatly after it’s 0.
As I’ve mentioned it already : On a different system with an PSU that definitely fulfills the PSU requirements.
Ask a friend, if you know one that has a similiar system.
Can you run this in the terminal and post the output, please?
bash -c 'for f in /sys/class/drm/card*/device/hwmon/hwmon*/power*; do echo "$f:"; cat "$f"; echo; done'
/sys/class/drm/card1/device/hwmon/hwmon10/power: cat:
/sys/class/drm/card1/device/hwmon/hwmon10/power: Is a directory
/sys/class/drm/card1/device/hwmon/hwmon10/power1_average: 6000000
/sys/class/drm/card1/device/hwmon/hwmon10/power1_cap: 0
/sys/class/drm/card1/device/hwmon/hwmon10/power1_cap_default: 0
/sys/class/drm/card1/device/hwmon/hwmon10/power1_cap_max: 350000000
/sys/class/drm/card1/device/hwmon/hwmon10/power1_cap_min: 0
/sys/class/drm/card1/device/hwmon/hwmon10/power1_label: PPT
So for some reason the power1_cap and pwer1_cap_default is set to 0 Watt instead to the expected/sensible values. power_cap_max with 350 Watt is still there.
Have to be away from keyboard for now, but I it were my card I would look into that: Why is it showing proper values on the live iso? Where are these values coming from? Can you change them? Maybe something useful in the bug trackers?
I did try echoing in some values yesterday but I was worried about doing damage. Also, isn’t LACT doing just that?
Not familiar enough with LACT and card tuning, usually just run default. Seems reasonable to assume it does.
I would look out more for: If the card report these unusual values by default without force setting anything, could that be an indicator for something else being wrong.
In case you haven’t disabled the lactd daemon via systemctl it might be still doing that.
Yes, disabled it this afternoon.
you can try occt in AUR
But what? I’m starting to suspect my screen since I get this when forcing VRR (or whatever is “Adaptive sync” in Plasma settings). Testing that because I user had similar problems that were solved by forcing it on.
You’re guessing. And that won’t address your issue. As @ricklinux said: You’re wasting your time.
You’re connected via HDMI. Adaptive sync would require HDMI 2.1 with VRR support. And I strongly doubt it that an older AOC 24" with a refresh rate of 75 Hz would have an HDMI 2.1 port, maybe HDMI 1.4 or so.
And that wouldn’t have any effect on the GPUs performance.
It’s working on Arch, thanks again to everybody who tried to help. I’ll see if I can find what’s causing the problem on EOS (after some rest).
I find this strange. My cards work fine on eos. I have an RX 6750XT also besides the RX 590 which would undoubtedly get way more FPS. Is Arch installed on the same computer? Or is this with the Gpu installed in a different computer with Arch?
Same PC, different disk, tried EOS on that too.
The math isn’t mathing! Doesn’t make any sense. I’m sure if i stuck that card in my system on my eos install it would work.
I agree, NUCs aren’t very popular, Windows 11 install doesn’t even have wifi drivers. It’s probably some BIOS settings or MB peculiarity.
But it gets to 420 too 
EDIT: I won’t ship it to you for testing, sorry.
Sorry I can’t come and get it either. I guess your stuck with the dilemma.
Edit: Are you running this on X11 or Wayland? On Arch and on EOS?
Wayland on both. EDIT: I think, can you choose that on EOS? I went with default.