Closed xM8WVqaG closed 2 years ago
Hi @xM8WVqaG,
The information that shows the global GPU usage is extracted from the Linux kernel DRM interface.
The per-process GPU usage is independent and is retrieved from /proc/
Your GPU being from 2013 I suspect that the DRM interface was not updated while the second one was.
Could you please try the patch (29e0d6968b9694dae9854b8cd196d639ee4e30f0) I pushed to the gpuUsageOldAMD branch to see if it fixes your issue?
To do so:
cd nvtop
git fetch origin
git checkout gpuUsageOldAMD
and follow the compilation process from the README.
Fantastic, thanks for the quick turn around! I can confirm your solution means GPU% is now tracked on my hardware.
Important Information
nvtop
master (123908f)Issue
At the top: GPU is reported as
N/A
, but MEM correctly updates. Plot: GPU% is always0
, but MEM% correctly updates. Table: GPU% updates on individual processes just like MEM% and CPU% values.See screenshot:
It's unexpected that
nvtop
is inconsistently reporting GPU metrics.