SOLVED: GPU support broken in 1.32 container.
Posted: Mon Jul 26, 2021 12:19 pm
Hi!
I try to use an nvidia quadro card for hw accelerated decoding.
Installed the nvidia docker toolkit on the host, using docker compose with these options:
runtime: nvidia
deploy:
resources:
reservations:
devices:
- capabilities: [gpu]
environment:
- "NVIDIA_VISIBLE_DEVICES=all"
- "NVIDIA_DRIVER_CAPABILITIES=all"
In the container the nvidia-smi shows this:
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.42.01 Driver Version: 470.42.01 CUDA Version: 11.4 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Quadro P600 On | 00000000:0A:00.0 Off | N/A |
| 34% 36C P8 N/A / N/A | 1MiB / 1999MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
So the container sees the GPU and its working.
ffmpeg shows this:
Hardware acceleration methods:
vdpau
vaapi
cuvid
So i trying to use 'cuvid' or 'vdpau'
I try to add to the ffmpeg parameter on the settings, image settings the ffmpeg hwaccel paramter, and using cuvid on the monitor settings page where the hwaccel name and device is.
The monitor up and running no error, but still not using the GPU, its using the CPU. I try with hwaccel device name: cuvid
hwaccel device: /dev/nvidia0 or just '0', or leave it empty. No difference.
Anybody can make this work anyhow? Im spent 2 days without success. I installed zm on the host so that not using docker, same issue. hw acceleration not working neither.
Thanks, Peter
I try to use an nvidia quadro card for hw accelerated decoding.
Installed the nvidia docker toolkit on the host, using docker compose with these options:
runtime: nvidia
deploy:
resources:
reservations:
devices:
- capabilities: [gpu]
environment:
- "NVIDIA_VISIBLE_DEVICES=all"
- "NVIDIA_DRIVER_CAPABILITIES=all"
In the container the nvidia-smi shows this:
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 470.42.01 Driver Version: 470.42.01 CUDA Version: 11.4 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Quadro P600 On | 00000000:0A:00.0 Off | N/A |
| 34% 36C P8 N/A / N/A | 1MiB / 1999MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
So the container sees the GPU and its working.
ffmpeg shows this:
Hardware acceleration methods:
vdpau
vaapi
cuvid
So i trying to use 'cuvid' or 'vdpau'
I try to add to the ffmpeg parameter on the settings, image settings the ffmpeg hwaccel paramter, and using cuvid on the monitor settings page where the hwaccel name and device is.
The monitor up and running no error, but still not using the GPU, its using the CPU. I try with hwaccel device name: cuvid
hwaccel device: /dev/nvidia0 or just '0', or leave it empty. No difference.
Anybody can make this work anyhow? Im spent 2 days without success. I installed zm on the host so that not using docker, same issue. hw acceleration not working neither.
Thanks, Peter