https://bugs.kde.org/show_bug.cgi?id=525406

--- Comment #2 from [email protected] ---
Additional diagnostic information:

The issue is still reproducible after updating to Plasma 6.7.5.

System information:
Operating System: CachyOS Linux
KDE Plasma Version: 6.7.5
KDE Frameworks Version: 6.30.0
Qt Version: 6.11.2
Kernel Version: 7.2.4-1-cachyos (64-bit)
Graphics Platform: Wayland
GPU: NVIDIA GeForce RTX 5070 Ti
NVIDIA driver: 615.71.09

The system also contains the integrated AMD Granite Ridge GPU, but the displays
and the Plasma Wayland session are running on the NVIDIA GPU.

The issue is reproducible with the NVIDIA GPU temperature sensor by itself. CPU
temperature sensors scale correctly, but adding the NVIDIA temperature sensor,
either alone or together with CPU sensors, causes the automatic Y-axis range to
become 0-5°C.

Disabling Automatic Y Data Range and setting a manual range, for example
0-100°C, makes the graph render correctly. This confirms that the temperature
samples themselves can be plotted correctly and that the problem specifically
affects the automatic range calculation.

The relevant NVIDIA temperature information reported by `nvidia-smi -q` is:

GPU Current Temp: 36 C
GPU Current T.Limit Temp: 52 C
GPU Shutdown T.Limit Temp Specification: -5 C
GPU Slowdown T.Limit Temp Specification: -2 C
GPU Max Operating T.Limit Temp Specification: 0 C

The current temperature is valid, but some temperature-limit metadata is
reported as -5°C, -2°C, and 0°C. I do not know whether the widget uses any of
these fields when calculating its automatic range, but the reported 0°C maximum
operating limit may be relevant to the resulting 0-5°C Y-axis.

For comparison, querying the current value directly also returns a valid
temperature:

`nvidia-smi --query-gpu=temperature.gpu --format=csv`
temperature.gpu
42

The full `nvidia-smi -q` output is available if needed.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to