Advanced Seminar at RCS
In the advanced seminar, you are going to study some of the recent topics in the fields of real-time and embedded systems. The seminar gives you the opportunity to practice working in a scientific manner. Therefore, you will need to research a given topic, write a paper, present it and write reviews on papers written by your fellow students.
Topics in winter term 2017/18
In the upcoming winter term, this seminar will cover the ongoing convergence of High-Performance Computing (HPC) architectures and applications with traditional, resource-constrained Embedded and Real-Time Systems.
Fore more details please check the list of topics below.
Mentoring and Registration
Each topic is supervised by a member of the chairs's research staff. Coordination and contact person for the winter term is Martin Geier.
To register to the course, please send an e-mail including
- Your name,
- Your matriculation number and
- The chosen topic.
Direct registration via TUM-Online is not possible to students. The registration is binding!
This winter semester presentations and reports will be accepted only in English.
- The (mandatory) kick-off meeting takes place on Thursday, 26.10.2017, at 13:15 in room 3999
- The actual seminar event is currently scheduled for Thursday, 01.02.2018, but can be adjusted at the kick-off
The seminar will cover the following topics:
- The drift of Applications - From clusters to smartphones [assigned]: More and more applications formerly only suitable for large distributed architectures can now be executed on modern embedded systems due to their continuously increasing processing power. This topic introduces such applications from domains like, e.g., Physics, Medical and Classification.
- Libraries and Tools for small and large high-performance systems: This topic explores tools (e.g., Vivado HLS) and libraries (e.g., OpenMPI) for communication and computation on such platforms.
- Interconnects and Networking: This topic covers the variety of (inter-node) communication systems ranging from Infiniband to IP-based networks.
- Node Architectures [assigned]: This topic explores the evolution of individual compute nodes from single-processor to heterogeneous multi-core systems (e.g., big.LITTLE) and their associated (intra-node) interconnects.
- The evolution of GPUs [assigned]: This topic explores the evolution of GPUs from gaming devices to GPGPU accelerators in HPC and embedded systems.
- The role of FPGAs, DSPs and ASICs in HPC and Embedded [assigned]: This topic examines the role of programmable, special-purpose hardware ranging from CPU accelerator interfaces (e.g., IBM CAPI) to ASICs such as Googles TPU.
Expected Number of Participants
The advanced seminar will be offered for a maximum of six students.