Server

During test drives, tests and competitions, all data generated in the vehicle by sensors and software are recorded.
Certain situations can be analyzed and replayed in the office.
This enables us to continuously improve our algorithms based on realistic data.

This results in very large amounts of data that have to be stored and managed.
For this purpose we have a server with 100TB capacity to which every employee has access.

 

In addition to the archive server task, other services also run on the server.

• Univention UCS for user administration and domain controllers
• Gitlab as web-based git repository, wiki and ticket system

CUDA

GPUs are highly scalable processor architectures with potentially thousands of cores and, compared to CPUs, magnitudes higher memory bandwidth.
On the other hand, they have less single-thread performance.
Therefore, GPUs are predestined to process the scalable part of the code, whereas the CPU processes the other part.
For example, the training and inference of Deep Neural Networks benefits from the use of GPUs.
For this purpose, we have our institute's CUDA-cluster. The dominant programming language for Nvidia-GPUs is CUDA.
In the following a fraction of our institute's applications:

• SIFT-Features
• Stereo through Semi-Global-Matching
• Pixelwise segmentation to enhance our online calibration

Continuous Integration (Build-Server)

Software quality plays a decisive role in development. We are all working on a common code base to benefit from each other's developments.

For this purpose, our institute has a build server with continuous integration. The developed code is only integrated if it passes all tests successfully.
This ensures that all employees always have access to a functioning software version.