Conical Systolic Array for Matrix Inversion

Matrices have been used in many analytical and simulation models and numerical solutions. Matrix operations have essential role in many scientific and engineering applications. One of the most time-consuming operations among matrix operations is matrix inversion. Many hardware designs and software algorithms have been proposed to reduce the time of computation. They will be more important for the large size matrix. In this paper a systolic array structure is proposed for matrix inversion, which is called Conical Systolic Array, CSA. Gaussian elimination algorithm is used in the CSA. Comparing previous designs, the number of processing elements, PEs, for CSA is 25% less than the previous design yet the total computation time is lower than the previous array.

[1]  J. Douglas Faires,et al.  Numerical Analysis , 1981 .

[2]  Emilio L. Zapata,et al.  Direct mapped cache performance modeling for sparse matrix operations , 1999, Proceedings of the Seventh Euromicro Workshop on Parallel and Distributed Processing. PDP'99.

[3]  Sang Uk Lee,et al.  Models and algorithms for efficient color image indexing , 1997 .

[4]  Roland W. Freund,et al.  Reduced-order modeling of large linear passive multi-terminal circuits using matrix-Pade approximation , 1998, Proceedings Design, Automation and Test in Europe.

[6]  Hadi Shahriar Shahhoseini,et al.  A new matrix method for pulse train identification: Implementing by systolic array , 2002, 2002 11th European Signal Processing Conference.

[7]  Ahmed El-Amawy.  A Systolic Architecture for Fast Dense Matrix Inversion , 1989, IEEE Trans. Computers.

[8]  Madhavan Swaminathan,et al.  Modeling of multi-layered power distribution planes including via effects using transmission matrix method , 2002, Proceedings of ASP-DAC/VLSI Design 2002. 7th Asia and South Pacific Design Automation Conference and 15h International Conference on VLSI Design.

[9]  P. Sadayappan,et al.  Circuit Simulation on Shared-Memory Multiprocessors , 1988, IEEE Trans. Computers.

[10]  Graham M. Megson.  Introduction to systolic algorithm design , 1992 .

[11]  Volker Zerbe,et al.  Parallel matrix computations and their applications for biomagnetic fields , 1997, Proceedings. Advances in Parallel and Distributed Computing.