Hasso-Plattner-InstitutSDG am HPI
Hasso-Plattner-InstitutDSG am HPI

Energiebewusstes Rechnen in hetrogenen Rechenzentren (Wintersemester 2019/2020)

Dozent: Prof. Dr. Andreas Polze (Betriebssysteme und Middleware) , Sven Köhler (Betriebssysteme und Middleware) , Max Plauth (Betriebssysteme und Middleware) , Lukas Wenzel (Betriebssysteme und Middleware)
Website zum Kurs: https://www.dcl.hpi.uni-potsdam.de/teaching/energy19/

Allgemeine Information

  • Semesterwochenstunden: 4
  • ECTS: 6
  • Benotet: Ja
  • Einschreibefrist: 30.10.2019
  • Lehrform:
  • Belegungsart: Wahlpflichtmodul
  • Lehrsprache: Englisch

Studiengänge, Modulgruppen & Module

Data Engineering MA
IT-Systems Engineering MA
  • IT-Systems Engineering
    • HPI-ITSE-A Analyse
  • IT-Systems Engineering
    • HPI-ITSE-E Entwurf
  • IT-Systems Engineering
    • HPI-ITSE-K Konstruktion
  • IT-Systems Engineering
    • HPI-ITSE-M Maintenance
  • OSIS: Operating Systems & Information Systems Technology
    • HPI-OSIS-K Konzepte und Methoden
  • OSIS: Operating Systems & Information Systems Technology
    • HPI-OSIS-T Techniken und Werkzeuge
  • OSIS: Operating Systems & Information Systems Technology
    • HPI-OSIS-S Spezialisierung


The increasing amount of volatile renewable energy sources introduces a number of challenges for today's large-scale computing systems. To operate reliably and efficiently, computing systems must adhere not only to technical limits (i.e. thermal constraints) but they must also consider volatile characteristics of the electricity grid such as production, demand, and energy prices. Furthermore, it is always desirable to increase the energy efficiency of computing systems in order to reduce the ecological footprint as well as the operating costs.

With the increasing degree of heterogeneity in today's computing systems, the goal of this project seminar is to exploit varying hardware characteristics across CPU architectures (Intel/AMD/POWER/ARM), GPUs (NVIDIA, AMD, Intel, ARM), FPGAs (Xilinx), and possibly even other Accelerators to find the best possible trade-off between energy consumption and execution time for a given workload.

In this project seminar, you will learn how to port and optimize a given workload to one or multiple target platforms of your choice. Depending on your previous knowledge and your commitment, you are free to deepen your knowledge in familiar architectures, or you can use the opportunity to get acquainted with a more exotic platform.

The objective is to assess energy requirements and execution efficiency for a wide range of computing problems and help data center operators schedule workloads according to the energy market by the hour.


The participants are expected to have a fair knowledge of the C/C++ programming languages.

Prior experience with multithreaded CPU-programming, GPU programming, or VHDL as well as performance measurements on GNU/Linux systems, is useful but not a mandatory requirement.

You will received detailed introductions to the programming concepts and hardware architectures used.


You can work alone, or in a team up to 3 persons. All students/teams are required to meet regularly with their advisor to discuss progress and plan further steps.

At the end of the semester you will give a final presentation of your findings, as well as provide a short technical report for later reference.

The final grade will be determined from your project work throughout the semester, as well as your presentation and report.


In the beginning of the semester, kick-off and technical introductions will take place each

  • Tuesday, 09:15-10:45 in A-2.1

Note, that later throughout the semester no further introduction is required and you can use this slot to work on your projects.