Hochperformante Verarbeitung und Visualisierung von Massendaten
High Performance Processing and Visualization of High Volume Data
The HPVis research project is directed towards the development of a new method to improve the performance of standard PC applications by the use of algorithm implementation on FPGAs. The objective is to develop an effective method to use an FPGA board via the PCI Express interface of standard PC systems for algorithm acceleration. The project’s focus lies on industrial sensor systems and image processing systems with high-volume data.
The main application which is analyzed visualizes pipeline inspection measurements from the partner company Rosen TRC. During a pipeline inspection a huge amount of measurement values are stored. They must be analyzed and visualized on a standard PC after the measurement procedure. The new method that is developed shall improve the processing performance of the underlying algorithms.
Another application field that is considered in the project is image processing. The methods investigated for the pipeline inspection application will be applied to smart cameras containing FPGAs which are produced by the partner company Visiosens.
- High-Performance Processing and Visualization of High-Volume Data, presented at the 8th International Conference on High-Performance and Embedded Architectures and Compilers (HiPEAC 2013), Berlin, Germany, January 2013
- M. Weinhardt, A. Krieger, Th. Kinder: A Framework for PC Applications with Portable and Scalable FPGA Accelerators, Proceedings of the International Conference on ReConFigurable Computing and FPGAs (ReConFig 2013), Cancun, Mexiko, December 2013, © 2013 IEEE1
- M. Weinhardt, R. Höckmann, T. Kinder, High-Level Design of Portable and Scalable FPGA Accelerators, First International Workshop on FPGAs for Software Programmers (FSP 2014), Munich, Germany, September 2014
- Meltebrink, Christian: Algorithmenbeschleunigung durch FPGAs zur Massendatenverarbeitung. Erschienen in INFORMATIK2014 - Big Data – Komplexität meistern (Lecture Notes in Informatics, Volume 232), S. 2493
- We presented our results during the demo night held at the 24th International Conference on Field Programmable Logic and Applications Munich, Germany; September 2 - 4, 2014