Making use of all hardware accelerators on Xavier and Orin

Hi All,

What would be best way/SDK to make use of hardware accelerators on Jetson Xavier and Orin (PVA, NVENC, NVDEC, OFA, DLA) ? Is DeepStream the way to go ? Would this be the best guide to use in order to install DeepStream SDK ?


If you need to run deep learning inference in your use-case, we suggest use DeepStream SDK. If your use-case is multimedia-only, we suggest use jetson_multimedia_api. Please check document of jetson_multimedia_api:
Jetson Linux API Reference: Main Page | NVIDIA Docs

Thanks DaneLLL. I need both. Need to decode (nvdec) and infer and re-encode (nvenc). I guess you would suggest to use DeepStream SDK. Is that right ?
How can I ensure that hardware accelerators are being used ?