I am currently using the R200 on the NVIDIA Jetson TX1 development board (and a multi camera setup in near future). Since I have tons of GPU power available it would be nice to actually have the camera driver (librealsense and the ros-wrapper) running on the GPU instead of the CPU.
However, before even thinking about it - what does librealsense actually do? I mean is it just handling the communication between camera and CPU or is actual image processing performed? If the latter is true then it might be nice to move some of the functions to the GPU.
Any inputs folks?
Hello Bob, Librealsense provides drivers that transfers the stream data from the camera through the CPU and up the OS stack and it does stream processing so that you are not just seeing raw data. You can find all information on librealsense here: https://github.com/IntelRealSense/librealsense.
The camera drivers all run on Intel CPUs only.
I hope this answers your question. If you would like further assistance, please post your question in the RealSense Community. This forum in the Intel Developer Zone will soon be archived and all support will be handled in RealSense Community.
Intel Customer Support