Technical Background

The NVIDIA Jetson TK1 is a single board computer with an ARM-based central processing unit (CPU). It’s a quad-core CPU based on the ARM Cortex A15 architecture. It further integrates a graphics processing unit (GPU) based on the NVIDIA Kepler architecture.

The NVIDIA Jetson TK1 supports Gigabit Ethernet with Jumbo Frames and data rates up to 1 Gbit/sec as well as USB 3.0 featuring data rates of up to 5 GBit/sec.

Getting Started

Installing a new OS on NVIDIA Jetson TK1

Installing a new OS on NVIDIA Jetson TK1 requires a host PC with Linux OS (we recommend using Ubuntu 14.04). It is necessary to install the OS version ”Linux for Tegra R21.5” on the NVIDIA Jetson TK1 to get the Baumer SDK working.

First download the Driver Package and the Sample File System from the following link:


Then, follow the guide on the NVIDIA website to install the new OS:



Please refer to the Installation Guide for Linux on how to connect a camera, aquire images and build
an application.


Installing Baumer GAPI SDK

To install the Baumer GAPI SDK download the NVIDIA Jetson TK1 software package from the Baumer website ( You can download the software package directly to the NVIDIA Jetson TK1 by connecting it to the internet. Then install the software by using the Linux terminal. Change to the directory where the downloaded file is stored (e.g. /home/user/Downloads).

The file is called “baumer-gapi-sdk-linux-v2.4-xxxxx-Linux4TegraR21.4-JetsonTK1.deb” [1].

Start the installation by executing the following command:


# sudo dpkg -i baumer-gapi-sdk-linux-v2.4-xxxxx-Linux4TegraR21.4-JetsonTK1.deb1

[1] xxxx is a version number




CPU Usage and Power Consumption

Frame Rate

Data Rate

Power Consumption

CPU Usage**

5 fps

100 Mbit/s

6.9 W

11 %

53 fps

988 Mbit/s

7.4 W

15 %

107 fps

1931 Mbit/s*

8.8 W

20 %

Table 1: CPU Usage and Power Consumption


Table 1 shows the CPU usage and power consumption of the NVIDIA Jetson TK1 during constantly getting pictures from different Baumer GigE industrial camera. For that the Baumer SDK example 001 (image capturing by polling) with an increased amount of requested images was used.


Measurements progressed with optimized network configuration and hardware.


*) To operate with Dual GigE cameras the NVIDIA Jetson TK1 was expanded with a mini PCIe card which provides two additional GigE interfaces.

**) The Tegra K1 is able to deactivate cores depending on the load. To get a comparable CPU Usage value all cores were forced to be active. This also increases the Power Consumption, what means that it’s possible that the Power Consumption is lower in convenient use.




Please contact our Technical & Application Support Center with any questions.

Phone: +49 3528 4386 845
E-mail: [email protected]

To the top