Embarrassingly parallel algorithms for algebraic number arithmetic — and some less trivial issues
Representing algebraic numbers α and β by their defining polynomials is an alternative to the older representation in which the sum α + β would be represented by a list of character strings recursively involving root indices, +, −, x, and integer terms or radicands. Algorithms for sums, products, etc., in the defining polynomial representation, are based on the theory of symmetric functions and have relatively efficient implementations using polynomial resultants. Even better algorithms use power sums rather than the coefficients of the defining polynomials; on massively parallel systems α × β executes in constant time, and computation of α + β is linear (or logarithmic, if enough processors are available) in mn, where m and n are the degrees of the defining polynomials and mn is the degree of the result. However, given polynomials of degree m and n, these algorithms require their power sums to order mn. The best known power-sum algorithm is based on Newton's identity, which may be treated as a linear recurrence, whose solution, conventionally understood to be of complexity O(mn), will dominate the time of the multiplication algorithm and significantly increase the addition time. A parallel algorithm, reducing the recurrence solution to O((log2n)(log2m)), is discussed.
Keywordsalgebraic numbers power sums linear recurrences
Unable to display preview. Download preview PDF.
- Borchardt, C. W.  Développements sur l'équation à l'aide de laquelle on détermine les inégalités séculaires du mouvement des planètes. Jour. de Math. (Liouville) 12:50–67.Google Scholar
- Hirsch, M.  Hirsch's Collection of Examples, Formulae, and Calculations of literal Calculus and Algebra. Translated from the German, by the Rev. J. A. Ross, A. M. London: Black, Young, & Young (1st German edition, Berlin: Duncker und Humblot, 1804.)Google Scholar
- Kogge, P.  Parallel solution of recurrence problems. IBM Jour. of Res. and Dev. 18(1974):138–148.Google Scholar
- Loos, R. and G. E. Collins  Resultant algorithms for exact arithmetic on algebraic numbers. SIAM Rev. Jan. 1974:130–131.Google Scholar
- MacDuffee, C. C.  The discriminant matrices of a linear associative algebra. Annals of Math., 2nd Ser. 32(1931):60–66.Google Scholar
- Ponder, C. G.  Evaluation of “Performance Enhancements” in algebraic manipulation systems (thesis abstract). In Computer Algebra and Parallelism (Ed. J. Della Dora and J. Fitch), Academic Press, 1989, pp. 51–73.Google Scholar
- Selmer, E. S. [1966a] Linear recurrence relations over finite fields.(mimeograph) Department of Mathematics, Univ. of Bergen, Norway, 212 pp.Google Scholar
- Selmer, E. S. [1966b] On Newton's equations for the power sums. BIT 6(1966):158–160.Google Scholar
- Siebert-Roch, F. and J.-M. Muller  VLSI manipulation of polynomials. In Computer Algebra and Parallelism (Ed. J. Della Dora and J. Fitch), Academic Press, 1989, pp. 233–256.Google Scholar
- Weeks, D.  Formal representations for algebraic numbers. SIGSAM Bulletin #31 (ACM) (August 1974): 91–95.Google Scholar