Im trying to develop a sim that involves render products from 4 cameras in parallel. I’m then publishing semantic_segmentation data from this to ROS.
I notice that the simulation fps drops from 35 to about 5-8 when I activate the render products. Is this drop in performance normal? Should I expect that to slow down that much?
Are there more clever ways to approach performance with applications that involve multiple cameras?
Im on isaac 4.2, everything is running in a docker container.