EU e-Privacy Directive

This website uses cookies to manage authentication, navigation, and other functions. By using our website, you agree that we can place these types of cookies on your device.

You have declined cookies. This decision can be reversed.

Research at the Institute for Computer Engineering

On this page we briefly introduce the chairs at the Institute for Computer Engineering of the University of Heidelberg and their research areas:


Application Specific Computing

Our research focuses on significant improvements of performance and accuracy in application specific computing through a global optimization across the entire spectrum of numerical methods, algorithm design, software implementation and hardware acceleration.

These layers typically have contradictory requirements and their integration poses many challenges. For example, numerically superior methods expose little parallelism, bandwidth efficient algorithms convolve the processing of space and time into unmanageable software patterns, high level language abstractions create data layout and composition barriers, and high performance on today's hardware poses strict requirements on parallel execution and data access. High performance and accuracy for the entire application can only be achieved by balancing these requirements across all layers.

The following topics are given particular attention:
- Mixed precision methods
- Multigrid methods
- Adaptive data structures
- Data representation
- Bandwidth optimization
- Reconfigurable computing 

Visit the website

top icon


Computer Architecture

CAG logo

The Computer Architecture Group at the University of Heidelberg has the expertise to design complex hardware/software systems. As system architects we cover not only the operation principles but include the technology and the software to build real working prototypes. All levels of system design are covered, starting at the application programming interface, e.g. MPI, through the efficient design of device drivers finishing at custom build hardware devices based on standard cells.

The group mainly focuses on the design of parallel architectures which achieve their high performance by improving communication between computational devices/units. Scaling such systems is a great challenge to the architecture of the interconnection network (IN) and the network interface controller (NIC). Special attention is paid on the interface between software and hardware to setup communication instructions.


Areas of Research:

  • Parallel- and cluster computing
  • New computer architectures und paralell computers
  • Low latency interconnection networks
  • Communication in cluster systems
  • Efficient hardware development tools

Projects include:

  • Center of Excellence for Hypertransport
  • HTX Board: A universal Hypertransport test platform
  • SEED: Support for Education in Electronic Design
  • ATOLL: Atomic low latency interconnect for cluster computing
  • FSM Designer 4: An efficient FSM design tool

Visit the website

top icon


Computing Systems

The Computing Systems Group is most concerned with computational growth under hard power constraints, and focusses on the interface of computer architecture and application. While Moore’s law continues to provide us CMOS chips with an increasing amount of transistors, we can no longer maintain a constant power consumption per component (cf. Fig, right). As a result, we have been witnessing the rise of multi- and many-core processors. While using such architectures has helped to maintain computational growth, the resulting heterogeneity has huge implications on the complexity of programming. Our research focusses on performance, energy efficiency and programmability of computer systems, ranging from small mobile devices to large datacenters, for applications including deep learning, high-performance computing and high-performance analytics.


In the GPU Mekong research project, we are addressing the issues of programmability for multi-GPU systems (cf. Fig, left). As a result, the programmer can see a multi-GPU system as a large single GPU, which substantially reduce programming complexity. Key of GPU Mekong is the ability to automatically, i.e., without any user intervention, create data movements to solve dependencies, which is usually a tedious and error-prone task of the programmer. This research effort has received a Google Faculty Research Award in 2014, and we collaborate with the EMCL group, ETH Zürich and NVIDIA Germany.

In the DeepChip project (cf. Fig, right), we collaborate with Graz University of Technology on deep learning techniques for small mobile systems. We use techniques including reduced precision, sparsity and asynchrony to significantly reduce the requirements of deep neural networks in terms of compute and memory to enable an inference on resource-constrained systems without a loss in accuracy. As a result, deep learning techniques are no longer confined to large datacenters, but instead classifications and regressions can be performed directly on the device.

Other research efforts include graph computations on columnar in-memory databases (collaboration with SAP), data acquisition systems for high-energy physics experiments (collaboration with CERN), and understanding and optimizing the energy proportionality of scalable computer systems. We are most interested in applications with challenging demands, and explore how we can maintain computational growth without loss of programmability or energy efficiency. Previous collaborations include Georgia Tech, NVIDIA, AMD, SUN, and Technical University of Valencia. We are constantly searching for new undergraduate and graduate students for various positions, currently in particular in the area of Deep Learning and Advanced Compilation Techniques.

Visit the website

top icon


Circuit Design

At the chair of Circuit Design, microelectronic circuits are developed, tested and applied. These microchips often contain extremely sensitive, low noise amplifiers for capturing sensor data and modules for further analog and digital signal processing. The crucial parts of such chips are designed completely manually. They are simulated on the analog level to achieve a maximal performance. The designs are fabricated in state-of-the-art CMOS technologies and are put into operation here at the group. A typical use case consists not only of designing the chip, but also includes building suitable control and data acquisition systems, the control and synchronisation of all components and the analysis of the measured data.

Here are some examples of our recent chip developments:

  • Highly integrated circuits for positron emission tomography 
  • Readout electronics for DEPFET sensors for the future ILC detector
  • Chips for detecting X-rays with hybrid pixel sensors 
  • Novel monolithic pixel sensors 
  • Development of front-end electronics for the CBM experiment at FAIR at the GSI.
  • High-speed microscopy within the Viroquant project
  • Detectors for synchrotron experiments at DESY, ESRF and the future XFEL
  • Circuit design techniques for generation of secret keys for cryptography 

Without such chips and systems being highly optimised for special tasks many research projects today could not be realized.

For students the "Computer Architecture" and "Circuit Design" groups offer the specialisation course on "Chip Design". This starts with the transistor and the fabrication process technologies of chips and provides basic knowledge of analog design and circuits. The description and design of digital circuits is covered in detail and every step from the idea to a complete chip is practically exercised in lab courses. 

Visit the website

top icon


Optimization, Robotics & Biomechanics

The research focus of the Optimization, Robotics & Biomechanics (ORB) Chair is on modeling, optimization and simulation of dynamic motions of anthropomorphic systems, i.e. humans, humanoid robots, and virtual human characters. Other research directions include motions of industrial robots, art robots, flying robots and swarm robots. From a mathematical perspective, we are particularly interested in the application and development of efficient numerical optimal control, inverse optimal control and non-smooth optimization techniques for complex hybrid dynamical system models. We also develop efficient tools to set up realistic dynamical optimization models of humans, robots and other technical devices including rigid multibody system models, muscle models and neural control.

Our research projects cover the following topics:

  • optimization of humanoid walking motions in different terrains
  • human movement understanding & identification of underlying objective functions of human motions in different situations
  • generation of fast human-like walking, running, jumping, diving and other gymnastics motions stability optimization of human and robot motions study of
  • characteristics of pathological gait in orthopedics and of walking motions with prostheses, orthoses and functional electrical stimulation
  • optimization of the design and control of exoskeletons
  • optimization of physically assistive devices for the elderly
  • trajectory optimization for robots
  • studies of artistic and emotional aspects of dynamic motions and development of art robots
  • investigation of processes related to cognition and orientation during locomotion and traffic interaction
  • controlling octocopters for automated photogrammetric reconstruction in archeology
  • needle path planning in robot assisted prostate brachytherapy and development of training environments
  • optimal control studies of manipulation combining motor control and biomechanical modeling approaches

Our interdisciplinary research creates bridges between scientific computing and many other disciplines, such as robotics, engineering, biomechanics, medicine, orthopedics, sports, computer graphics, cognitive sciences, art and archeology.

Visit the website

back to top