Jetson AGX Xavier vs Quadro P3000 inference performance?


For a proof of concept project I want to do real-time inference on a camera stream on the edge and I am looking into my hardware options. Biggest constrain is the maximum allowed power draw. My options thus far are the Jetson AGX Xavier or a x86 platform with a Quadro P3000. I assume the P3000 is more powerful given the higher CUDA core count, but the Xavier is of a newer generation and thus offers better FF16 performance. I’m having trouble finding good inference comparisons between these cards. Does anyone know the performance difference between a Xavier and a P3000 (or other card from that generation)? And are there other things to take into account between these two platforms?


Sorry that we don’t have a comparison table for Xavier vs. Quadro card.
But you can find Xavier AI performance on the below page:

And detailed performance(ex. encoding/decoding) below: