Running Live Transcoder without GPU
If your instance doesn't feature Nvidia GPU, you may still run Live Transcoder with pipelines which don't involve GPU processing.
The following use cases can be used without a GPU:
Use case | Input | Output |
---|---|---|
MPEG-TS UDP, RTP, SRT - Video passthrough | H.262, H.264, H.265, JPEG2000, JPEG-XS | H.262, H.264, H.265, JPEG2000, JPEG-XS |
MPEG-TS UDP, RTP, SRT - Framerate conversion | H.262, H.264, JPEG-XS | H.262, H.264, JPEG-XS |
Recommended instance type for AWS EC2 c5a.16xlarge
To enable CPU-only run, the docker-compose.yml must be changed this way:
- Open docker-compose.yml with a text editor:
vim docker-compose.yml
or
nano docker-compose.yml
- Delete or comment out (
#
) the following lines:
deploy:
resources:
reservations:
devices:
- driver: nvidia
count: all
capabilities: [gpu]
and:
- NVIDIA_VISIBLE_DEVICES=all
- Save the file (in the vim editor press
<esc>
key, then:wg
+<enter>
; in the nano editor press<ctrl>
+X
and confirm by pressingY
, then<enter>
)
Updated 14 days ago