Gustav – rebotnix technologies

GUSTAV offers reliably visual computing

Get real-time visual computing Artificial Intelligence (AI) performance where you need it most with the high-performance, low-power by hundreds of GPU Cores, Fan-less and Black Anodized Alumimium. Gustav is the fastest AI supercomputer, based on NVIDIA™ Jetson® TX2. Best of all, it packs this performance into a small, power-efficient form factor that’s ideal for intelligent edge devices like security, drones, cars and medical devices. With our partner Raypack.ai, we offer several software -packages for Gustav, based on Maschine-Learning / Deep learning.

1.5 TERAFLOPS OF PERFORMANCE

A powerful AI device

Gustav gets is equipped with an operating system. You get full access as root on the system. In addition to the current CUDA drivers, all hardware relevant drivers like USB, Gigabit or CAN-Bus or UART are supported. Software developers and companies can start immediately with their projects. Gustav is the perfect in-house (AI) solution.

We support the best Deep Learning Frameworks

TensorFlow is an open-source software library for dataflow programming across a range of tasks. It is a symbolic math library, and also used for machine learning applications such as neural networks. TensorRT is developed by Nvidia. Caffe is a deep learning framework, originally developed at UC Berkeley. It is open source, under a BSD license. It is written in C++, with a Python interface as TensorFlow.

Turn your old infrastructure into AI

Artificial intelligence only makes sense if it is cost-effective. Transform your old IP-based video infrastructure into a visual solution with artificial intelligence. Whether you’re using IP cameras or USB webcams, the USB 3.0 and Gigabit Ethernet interfaces makes GUSTAV the perfect choice.

GUSTAV is swarm compatible

Instead of a super computer which is able to process a limited number of video signals, you can add any number of GUSTAV´s to your network. You can start with one device and add as many devices as you like later on. The swarm can be assigned different AI tasks at intervals. For example, a group of cameras can be assigned to a group of Gustavs. The modular possibilities of a Swarm intelligence are unlimited for the first time.

Please note that the swarm distribution software is a special product of our partners which use GUSTAV as a base.

That's how we make Gustav.

Some Customer Applications

Realtime Object Segmentation

Realtime Classifications

Realtime Visual Office work

A.I. Car Applications

Faster then Realtime Video Transcoding

All customer applications shown are subject to confidentiality agreements with our customers. We will present some public examples soon.

Here are the technical details

GUSTAV is available in two versions. The Model A has 2 x USB 3.0, a Gigabit network interface and a power jack. The Model B also has 2 x USB 3.0, a Gigabit network interface and on the rear side 2 CAN busses and 1 UART bus. Model A and B models can also get an internal 256 Gigabyte SSD hard drive. The size does not change if you select the optional internet hard disk.

Datasheet

GUSTAV Modell A

NVIDIA™ Jetson® TX2

GPU NVIDIA Pascal™, 256 CUDA cores
32GB eMMC storage capacity
CPU HMP Dual Denver 2/2 MB L2 +Quad ARM® A57/2 MB L2
VIDEO 4K x 2K 60 Hz Encoding (HEVC)
4K x 2K 60 Hz Decoding (12-Bit Support)
MEMORY 8 GB 128 bit LPDDR4 59.7 GB/s

Mainboard

HDMI OUT (MINI) 2.0
2 x USB 3.0
1 x Gigabit Ethernet
12-17 Volts Input (Lemo Connector)
Length 12 cm (4.72 inch)
Width 8 cm (3.14 inch)
Height 4 cm (1.57 inch)
Nvidia Jetpack 3.1 & 3.2 Support

Addons

Addons for Modell A & B

250 Gigabytes SSD – Samsung 960 Series
500 Gigabytes SSD – Samsung 960 Series
1000 Gigabytes SSD – Samsung 960 Series

GUSTAV Modell B