Open yayaasd opened 1 month ago
Thank you for this suggestion, there are actually 3 separate notes I have to mention:
GPU-Util
is not a useful indicator for "GPU is bored", because Handbrake does use the NVEnc dedicated encoder on the GPU which is not included in the GPU-Util
metric in your screenshot.
As you can see in this screenshot, the ENC is 100% while GPU is actually at 0% load for ghb
.
This is because GPU-Util only measures the compute / graphics load of the CUDA cores and not the dedicated video encoder.
Given that your 3050 has only one encoder chip, I am pretty sure that it is fully loaded and running multiple encodes at the same time probably would not increase your encoding speed. source
Dear all, docker with nvenc support ist running and automatic conversion from watched folder is just fine. only missing how to convert multiple files at the same time?
the GPU is kind of bored ;-)
EDIT: Found some additional info: https://github.com/HandBrake/HandBrake/issues/1445