I have created an A* Pathplanning programm in C++, which can solve a planning problem.
During my testing I compared the performance of the programm and the actual planning is slower on the Jetson (as expected). It consists of some searching inside of a list.
The creation of the grid, in which I do the planning, seems to be way faster on the Jetson. It is doing a lot of Point-in-Polygon tests during the creation but as these are just normal calculations so I don’t understand the speed up.
I am not using CUDA or something like that which could maybe be an explanation for a faster computational time on the Jetson.
Did anyone experience something similar or has any idea on what could be the cause for that?
Significant Performance Difference in between High Performance Notebook and Jetson Xavier NX (in favor of Jetson)
What’s the system setup of your High Performance Notebook?
Have you maximized the the Jetson performance by NVPModel? See NVIDIA Jetson Linux Driver Package Software Features : Clock Frequency and Power Management | NVIDIA Docs
One possible reason is that the searching task allocates certain large memory.
Since NX has limited resources, it might be slow especially some swap memory is used.
Which platform do you compare to?
To give a further suggestion, we will need a sample to reproduce this issue in our environment first.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.