Automatic error cumulation control

  • B. J. A. Hulshof
  • J. A. van Hulzen
Applications 2
Part of the Lecture Notes in Computer Science book series (LNCS, volume 174)

Abstract

Algorithmic methods are presented to perform a priori error analysis and error cumulation comtrol. The corresponding programs are implemented in REDUCE as an extension of Sasaki's multiple precision floating point arithmetic package. Ingredients for the method are some concepts of interval arithmetic and a slightly modified precision notion, in both absolute and relative sense. This allows to take second order effects into account and to consider errors as a combination of inevitable propagated errors, due to inaccurate input, and controlable generated errors, due to arithmetic operations. The error control allows to consider these operations as adjustable sources and results in a set of instructions for using the Sasaki-package such that the precision, dictated by the analysis, guarantees to limit error cumulation to admissible, user chosen error bounds.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. [1]
    Bauer, F.L.: Computational graphs and rounding error, SIAM J. Numer.Anal. 11, 87–96 (1974).Google Scholar
  2. [2]
    Hearn, A.C.: REDUCE user's manual. The Rand Corporation (1983).Google Scholar
  3. [3]
    Johnson, D.B., Miller, W., Minnihan, B., Wrathall, C.: Reducibility among floating-point graphs, J. ACM 26, 739–760 (1979).Google Scholar
  4. [4]
    Kulish, U., Miranker, W.L.: Computer arithmetic in theory and practice. New York: Academic Press (1981).Google Scholar
  5. [5]
    Lanam, D.H.: An algebraic frond-end for the production and use of numerical programs, Proceedings SYMSAC '81 (P.S. Wang, ed.), 223–227. New York: ACM (1981).Google Scholar
  6. [6]
    Larsen, L., Sameh, A.: Efficient calculations of the effect of roundoff errors, ACM TOMS 4, 228–236 (1978).Google Scholar
  7. [7]
    Miller, W.: Software for roundoff analysis, ACM TOMS 1, 108–128 (1975).Google Scholar
  8. [8]
    Miller, W.: Spooner, D.: Software for roundoff analysis II, ACM TOMS 4, 369–387 (1978).Google Scholar
  9. [9]
    Olver, F.W.J.: A new approach to error arithmetic, SIAM, J. Numer.Anal. 15, 368–393 (1978).Google Scholar
  10. [10]
    Richman, P.L.: Automatic error analysis for determining precision, C. ACM 15, 813–817 (1972).Google Scholar
  11. [11]
    Sasaki, T.: An arbitrary precision real arithmetic package in REDUCE, Symbolic and Algebraic Computation (E.W. Ng, ed.), LNCS series nr.72, 358–368. Berlin-Heidelberg-New-York: Springer Verlag (1979).Google Scholar
  12. [12]
    Sterbenz, P.H.: Floating point computation. New York: Prentice Hall (1974).Google Scholar
  13. [13]
    Stummel, F.: Perturbation theory for evaluation algorithms of arithmetic expressions, Math. of Comp. 37, 435–473 (1981).Google Scholar
  14. [14]
    van Hulzen, J.A.: Code optimization of multivariate polynomial schemes: a pragmatic approach, Proceedings EUROCAL '83 (J.A. van Hulzen, ed.), LNCS series nr. 162, 268–300, Berlin-Heidelberg-New York: Springer Verlag (1983).Google Scholar
  15. [15]
    van Hulzen, J.A., Hulshof, B.J.A.: A code optimization package for REDUCE (in preparation).Google Scholar
  16. [16]
    Wang, P.S., Chang, T.Y.P., van Hulzen, J.A.: Code generation and optimization for finite element analysis (These Proceedings).Google Scholar

Copyright information

© Springer-Verlag 1984

Authors and Affiliations

  • B. J. A. Hulshof
    • 1
  • J. A. van Hulzen
    • 1
  1. 1.Department of Computer ScienceTwente University of TechnologyEnschedethe Netherlands

Personalised recommendations