DeepLearning - CPU+GPU in car

With everything drawing power from the board’s USB bus, CSI-2 connector, or GPIO pins, the actual power consumption of a fully equipped Nano running at maximum clock speed sits between 15 ~ 25 Watts.

Using power supplies with higher-than-5-volt outputs, can severely damage your Jetson Nano . There are no voltage regulators on the official carrier board

But going with 5V adapters with more than 4 Amps are highly recommended, because the Nano only takes as many Amps as it consumes. So as long as your power supplies are within the 5V range, you will be safe.

NVIDIA Jetson Neuron Board ( from Mistral Solutions )

How to measure power used by Edge computing?

As a first step, look for reliable ways to measure the power for the whole board as well as for single layers of a deep neural network by reviewing and testing some standard state-of-the-art methods. There are several methods available, e.g.,measuring the power of existing sensors or using available internal tools on the board. The challenge is to interpret the results as background processes are considered.

Back

Page updated

Report abuse