zur Startseite


Implementation of State-of-the-art Losless Coding on Parallel Systems (PPM)
Betreuer Dipl.-Inf. Simeon Wahl


Prediction by Partial Matching (PPM) is a state-of-the-art lossless data compression scheme that reaches its unperceived compression performance by efficiently exploiting dependencies in datastreams by using context information for data prediction. In spite of the excellent compression ratios, PPM algorithms are still not widely used due to their high computational requirements. The goal of this project will be to explore the parallelization potential of PPM family and to create a fast and efficient parallel implementation of PPM.
The platform for implementation can either be a nVidia GeForce GPU using CUDA architecture or a Xilinx FPGA platform. Latter can be programmed using VHDL, Xilinx System Generator (Matlab/Simulink), or ANSI C/C++/SystemC in combination with the Catapult C platform from Mentor Graphics.

In this project you will have the opportunity to:

  • Study state-of-the-Art Compression Algorithms
  • Design and develop algorithms from Prediction by Partial Match family of algorithms
  • Optimize your algorithms for compression of images and sci data


The project is to be conducted at the Chair for Parallel Systems at IPVS, in a modern computing environment providing access to the latest graphics hardware and development utilities. The scope of the project will be adjusted to the requirements and duration of the thesis or project that you would like to complete.

Start: as soon as possible!
Requirements: Excellent C programming knowledge!
Working language: German or English.


If you are interested to work on this topic, send an application email to:
Simeon Wahl (wahlsn@ipvs.uni-stuttgart.de)