Systolic designs for evaluating linear combinations of Chebyshev polynomials

  • Octav Brudaru
  • Graham M. Megson
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 817)


We consider here the design of 2D systolic arrays for evaluating linear combinations of Chebyshev polynomials, a0T0(x)+ ...+a n T n (x), approximating real functions of one real variable. A tree-like systolic array and a systolic rosette are proposed. They can achieve a 2 log2n−1 response time and an one clock tick throughput rate at an expense of n− 1 simple cells. They are asymptotically optimal with respect to bus bandwidth utilization. The systolic rosette has a center of symmetry and a very regular interconnection network. Some cvasi-linear and non-linear transformations are used to obtain these architectures.


Systolic arrays cvasi-linear/non-linear transformations Chebyshev polynomials 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. [BGB93]
    O. Brudaru, D. Galea, M. Bulea, Linear systolic arrays for computing Chebyshev polynomials, RR 93-22, Institute of Theoretical Informatics, Romanian Academy, Iasi, 1993.Google Scholar
  2. [BPVL93]
    V. Beiu, J. Peperstaete, J. Vanderwalle, R. Lauwereins, Close approximations of sigmoid functions by sum of steps for VLSI implementation of neural networks, The 9th Romanian Symposium on Computer Science'93, Iasi, November 12–13, 1993.Google Scholar
  3. [CDH91]
    G. Corbaz, J. Duprat, B. Hochet, J. M. Muller, Implementation of a VLSI polynomial evaluator for real time applications, RR 91-07, LIP, ENSL, 1991.Google Scholar
  4. [Cra84]
    R. Cramariuc, Conformal Transformations in Physics and Electrotechnics, Scientific and Encyclopedic Editing House, Bucharest, 1984.Google Scholar
  5. [DM88]
    J. Duprat, J.M. Muller, Evaluation of polynomials and elementary functions, RR 698-1, TIM3, IMAG, 1988.Google Scholar
  6. [DDR91]
    A. Darte, T. Risset, Y. Robert, Synthesizing systolic arrays:some recent developments, RR 91-09, LIP, ENSL, 1991.Google Scholar
  7. [EMB89]
    D. J. Evans, K. Margaritis, M. P. Bekakos, Systolic and holographic pyramidal softsystolic designs for successive matrix powers, Parallel Computing, 9 (1989), 373–384.CrossRefMathSciNetGoogle Scholar
  8. [MBC92]
    G. M. Megson, O. Brudaru, D. Comish, Systolic designs for Aitken's root finding methods, Parallel Computing 18 (1992), 415–429.CrossRefGoogle Scholar
  9. [Sch89]
    F. Scheid, Theory and Problems of Numerical Analysis, Schaum's Outline Series, McGraw-Hill, New York, 1989.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 1994

Authors and Affiliations

  • Octav Brudaru
    • 1
  • Graham M. Megson
    • 2
  1. 1.Seminarul Matematic ”A. Myller”Universitatea ”Al.I. Cuza” IaŞiIaŞiRomania
  2. 2.Department of Computing ScienceUniversity of Newcastle-Upon-TyneNewcastle-Upon-TyneUK

Personalised recommendations