Issues with Understanding Virtual Channel Limitation


I am working on scoping a project on the Xavier NX. We plan to utilize the system with a GMSL2 Camera setup.

I was curious if there is any documentation detailing the limitations of Virtual channels - ie if I run the maximum amount of channels, what sort of FPS can I expect from a 12 MP camera.

Just looking for something to read, there are not many details here: Jetson Virtual Channel with GMSL Camera Framework

You can calculate the throughput data if less than per lane 1.2G(DPHY).

Thanks Shane - can you validate my math here:

So if I had a 12MP camera at 10 bit images in RAW capturing at 60 fps (12 MP x 10 bit x 60 fps = 15 MB x 60 fps= 900 MBps)

If that then passes through a 4 x Aggregator (assume no loss here) with 4 cameras, I would be at (4 cameras x 900 MBps = 3.6 GBps) since this is across 2 lanes I would be at 1.8 GB per lane, so too much to run full frame rate.


Then working backwards from framerate I would be limited to 40 fps on each camera if I am just thinking about data.

Should be bit per seconds instead of byte per second for the calculation.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.