If you use cuda for nvr software
Will it put a lot of pressure on the cuda and gpu capabilities that can be utilized by the end user who purchased the nvr?
When making software using the jetson nano board, I know that you need to use the hardware features cuda and gpu in order to maximize performance.
However, if the nvr software consumes too much gpu resources, I think there may be a lot of restrictions on the resources that can be used when the buyer tries to do other work (video related) other than the nvr software.
How would you approach this problem?