search

NVIDIA / VideoProcessingFramework

Set of Python bindings to C++ libraries which provides full HW acceleration for video decoding, encoding and GPU-accelerated color space and pixel format conversions
Apache License 2.0
1.29k stars 231 forks source link

Does vpf now support jetson devices？ #523

Open 6xdax opened 11 months ago

6xdax commented 11 months ago

Is your feature request related to a problem? Please describe. install vpf in jetson devices

Describe the solution you'd like to decode and encode in jetson GPU

Describe alternatives you've considered opencv and Gstreamer or deepstream

Additional context Add any other context or screenshots about the feature request here.

RomanArzumanyan commented 10 months ago

Hi @6xdax

It does support jetson in feature branch which isn't merged to main branch yet.

6xdax commented 10 months ago

@RomanArzumanyan Thanks for your reply. But I can't find the branch. Can you tell me which branch it is? Thank you very much.

RomanArzumanyan commented 10 months ago

@6xdax

Since PR isn't merged yet, it's not in VPF but in clone repo, here's the MR: https://github.com/NVIDIA/VideoProcessingFramework/pull/515

6xdax commented 10 months ago

@RomanArzumanyan I have read it. And I have implemented jetson-ffmpegon Xavier NX board. It takes about 50% of one CPU to decode a 2560*1440@25fps rtsp video stream. But I don't know what the difference is between it and Jetson-ffmpeg. Whether it consumes less CPU resources?

RomanArzumanyan commented 10 months ago

@6xdax

But I don't know what the difference is between it and Jetson-ffmpeg. Whether it consumes less CPU resources?

That's a good question but it's outside the VPF scope. There's a VPF build for Tegra in MR, please check it out. If you're unhappy with it's performance - please LMK. You can use tegrastats utility on Xavier to see HW statistics.

6xdax commented 10 months ago

@RomanArzumanyan Thanks for your quick reply. I will try it.