Title :
Performance Evaluation of TriBA -- A Novel Scalable Architecture for High Performance Applications
Author :
Khan, Haroon-ur-Rashid ; Feng, Shi
Author_Institution :
Beijing Inst. of Technol., Beijing
fDate :
July 30 2007-Aug. 1 2007
Abstract :
This paper evaluates the performance of TriBA - a new idea in multiprocessor architectures. TriBA is considered to be a high performance, scalable parallel computing architecture, therefore, its performance is evaluated by well-known scalability metrics speed-up and efficiency. The root of "triplet based architecture" is based on the concept that "complex problems can be decomposed to three relatively independent sub-problems, which are data processing, data management and data communication". TriBA consists of a 2D grid of small, programmable processing units, each physically connected to its three neighbors. TriBA is a real scalable architecture, featuring fractal nature for computers. TriBA is a new solution for computer architecture, which is suitable for sophisticated embedded applications with multiple concurrent processing centers. The characteristics of this architecture are its great modularity, flexibility and scalability to meet the real-time signal processing demands in future telecommunication and multimedia systems. The feasibility, effectiveness and performance of TriBA is illustrated using execution model and simulations comparing it with conventional 2D mesh architecture.
Keywords :
multiprocessing systems; parallel architectures; performance evaluation; TriBA performance evaluation; computer architecture; data communication; data management; data processing; embedded application; high performance application; multiple concurrent processing center; multiprocessor architecture; programmable processing unit; real-time signal processing; scalability metrics; scalable parallel computing architecture; triplet based architecture; Application software; Computer architecture; Data communication; Data processing; Fractals; Parallel processing; Real time systems; Scalability; Signal processing; Telecommunication computing;
Conference_Titel :
Software Engineering, Artificial Intelligence, Networking, and Parallel/Distributed Computing, 2007. SNPD 2007. Eighth ACIS International Conference on
Conference_Location :
Qingdao
Print_ISBN :
978-0-7695-2909-7
DOI :
10.1109/SNPD.2007.411