Abstract
In this paper, effective algorithms and pipelined VLSI architectures are developed for the Cholesky factorization of a class of structured linear system equations. This method will be most effective for matrices which have low displacement ranks [1]. Our method is an efficient implementation of the generalized Schur algorithm [2]. Specifically, we prove that the generalized Schur rotation matrix can be decomposed into a sequence of elementary rotations and hence admits a simple and regular VLSI implementation using a linear array of doubly pipelined Cordic processors [3]. With a linear array of O(N) Cordic processors, it is able to solve a N-th order structured linear system with O[(α -1)N] time units.
Keywords
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
The second author is supported by National Science Foundation under contract ECS-840468
Preview
Unable to display preview. Download preview PDF.
References
T. Kailath, S.Y. Kung, and M. Morf," Displacement ranks of matrices and linear equations", J. Math. Anal. Appl., Vol.68, No.2, 1979, pp.395–407.
H. Lev-Ari, and T. Kailath, "Lattice filter parametrization and modeling of nonstationary processes", IEEE Trans. Information Theory, Vol. IT-30, No.1, Jan. 1984, pp.2–16.
Y.H. Hu," Pipelined Cordic architectures for the implementation of rotation-based algorithms", Tech. Rept. EE 85003, Electrical Engineering Dept., Southern Methodist University.
N. Levinson, "The Wiener RMS (root-mean-square) error criterion in filter design and prediction", J. Math. Phys., Vol.25, pp.261–278, Jan. 1947.
T. Kailath, "A view of three decades of linear filtering theory", IEEE Trans. Inform. Theory, Vol.IT-20, pp.145–181, Mar. 1974.
A.K. Jain, "Fast inversion of banded Toeplitz matrices by circular decomposition", IEEE Trans. Acoust., Speech Signal Processing, Vol.ASSP-26, pp.121–126, Apr. 1978.
M. Morf, "Doubling algorithm for Toeplitz and related equation", Proc. ICCASP 80, Denver,CO, Apr. 1980, pp.954–959.
S.Y. Kung and Y.H. Hu, "Fast and parallel algorithm for solving Toeplitz systems", Int'l Symp. on Mini and Micro-Computer in Control and Measurement, San Francisco, May 1981.
S.Y. Kung and Y.H. Hu, "A highly concurrent algorithm and pipelined architecture for solving Toeplitz systems", IEEE Trans. ASSP, Vol.ASSP-31, pp.66–76, Feb. 1983.
H.M. Ahmed, et al, "Highly concurrent computing structures for matrix arithmetic and signal processing", IEEE Computer, pp.65–82, Jan. 1982.
R.P. Brent and F.T. Luk," A systolic array for the linear time solution of Toeplitz systems of equations", J. of VLSI and Computer Systems, Vol. 1, No.1, pp.1–22.
J.M. Delosme and M Morf, "Scattering arrays for matrix computations", SPIE, Vol.298, Real-Time Signal Processing IV(1981), pp.74–83.
J.M. Delosme and I.C.F. Ipsen, "Efficient parallel solution of linear systems with hyperbolic rotations", Research Report YALEU/DCS/RR-341, Yale Univ., Nov. 1984.
S.L. Marple Jr., "High resolution autoregressive spectrum analysis using noise power cancellation", Proc. 1978 IEEE ICASSP, pp. 345–348.
B. Friedlander, "Lattice method for spectral estimation", Proceedings of the IEEE, Vol.70, Sept. 1982.
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 1986 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Jou, IC., Hu, YH., Parng, T.M. (1986). Vlsi algorithms and pipelined architectures for solving structured linear system. In: Makedon, F., Mehlhorn, K., Papatheodorou, T., Spirakis, P. (eds) VLSI Algorithms and Architectures. AWOC 1986. Lecture Notes in Computer Science, vol 227. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-16766-8_14
Download citation
DOI: https://doi.org/10.1007/3-540-16766-8_14
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-16766-2
Online ISBN: 978-3-540-38746-6
eBook Packages: Springer Book Archive