Run on laptop with nvidia GPU

[SuperElectron]

I am wondering if I can develop an application using this repo on my laptop, or if a board is required.

specs: nvidia GeForce GTX 1650 Ti Mobile, intel i7, ubuntu 20.04 os

Thanks in advance :)

[nhphuong91]

To put this simply, VVAS is a bunch of gstreamer (gst) plugins to accelerate gst video pipeline using Xilinx hardware. If you don't have any Xilinx hardware, then purchase one and try this. If not, you can just use the gst plugins which utilize GPU as hardware acceleration if any.

[SuperElectron]

Okay thanks for the reply!
One last thing, you mentioned ...:
If not, you can just use the gst plugins which utilize GPU as hardware acceleration if any.

Question: Can you give an example of "the gst plugins which use GPU hardware acceleration"?

1. My understanding is that your Xilinx inference you would run something like this (only on xilinx hardware)
filesrc -> qtdemux -> avdec_h264 -> ivas_xfilter -> ivas_metaaffixer -> ivas_xfilter -> videoconvert -> autovideosink

2. A substitute would be to use GPU accelerated elements (not on xilinx hardware, utilize GPU as hardware acceleration if any)
[?]

The context of my question was to gain an understanding if there is a development environment that does not require the hardware for teams to create code. In particular, I know that using DL Streamer or Deepstream requires practice and customization to get used of the metadata structure and lots of additional code is needed (so a development environment that doesn't need the hardware is great). Thanks in advance :)

[nhphuong91]

1. It's just as you mentioned, nvidia deepstream is a "nvidia version" of VVAS. Therefore, just look up any tutorials/examples using nvidia deepstream & you can start there.
2. About your concern, one solution could be to set up vscode remoted development environment on remote device (server or edge) & having your team develop remoting to it. To be concrete, it's not the same as remote to a server using RDP, instead, your team only need to install vscode locally & have it connecting to a remoted device which host a fully setup development environment. Refer to here for further info: https://code.visualstudio.com/docs/remote/remote-overview