I resently designed and created a new carrier board for Nvidia Jetson TX2 computer module. It worked well but there was a small problem that the loop time of the program I run in the Jetson TX2 shows very low performance from time to time. When I run the same program in the original development board, the looptime is nearly 3.0 milliseconds. But the loop time reaches to 4 secods when the program is run in the carrier board I designed. Can this be a design problem of my carrier board or is there any other reason?
This image shows the loop time vs, the program cycle count. It shows the loop time rises to 4 second for some program cycles.