Parallel Hardware Architectures for Computational Intensive and Secure Applications
Automatisierungstechnik, Mechatronik, Regelungssysteme, Intelligente Technische Systeme, Robotik
Elektronische Halbleiter, Bauelemente und Schaltungen, Integrierte Systeme, Sensorik, Theoretische Elektrotechnik
Zusammenfassung der Projektergebnisse
The main objective of this project was to explore new solutions for three important areas in computer science and electrical engineering: security, bioinformatics and autonomous mobile robotic systems. Partners from Universidade de Brasília (UnB) were providing the algorithms and experience from conventional hardware platforms whereas the Department of Electrical Engineering & Information Technology (ITIV) from Karlsruhe Institute of Technology (KIT) was providing different hardware platforms developed in the research projects Kahrisma, AMURHA, InvasIC and Aramis. As the Brazilian partners did not receive any funding, only the German partners were able to travel to Brazil. The German researchers have traveled to Brazil three times: The first two visits were focused on porting the particle swarm optimization (PSO) algorithm to the research processor Kahrisma from KIT. The performance was then compared with other softcore processors on FPGAs and hardware implementations. During the last visit, the German researchers held two workshops about the two projects InvasIC and Aramis. InvasIC is investigating a novel paradigm for the design and resource-aware programming of future parallel computing systems. Within the project, a whole framework including languages, compilers, operating systems and the hardware platform has been developed. The ARAMiS project is targeting multicore technologies for automotive, railway and avionic systems. During the workshop, the focus was put on using hardware accelerators on FPGA boards which can be accessed through the common PCI Express interfaces to speed-up the execution of applications on the host system. During the visits it became apparent that the different focuses of the institutes were well complementing each other. The researchers from ITIV have much experience with different kinds of parallel and reconfigurable hardware architectures as well as with tools to develop for these platform. Whereas the Brazilian researchers have much knowledge about different algorithms from security, bioinformatics and autonomous mobile robotic systems. The collaboration using the Kahrisma platform and the particle swarm optimization algorithms led to interesting results which can be used to choose the best fitting platforms for future systems. The workshops showed that this potential also exists in other projects but could unfortunately not be evaluated further during the project because of the lack of funding on the Brazilian side. The Brazilian partners were able to get funding for a follow-up project and could finally visit the researchers at KIT in August 2014. Here, they could bring in new impulses to the InvasIC and the ALMA projects.
