arXiv subjects

Richard P. Brent

Publications and source records attributed to Richard P. Brent.

65 records · Page 4Linked to original sources

arXiv2010

Modern Computer Arithmetic (version 0.5.1)

This is a draft of a book about algorithms for performing arithmetic, and their implementation on modern computers. We are concerned with software more than hardware - we do not cover computer architecture or the design of computer hardware. Instead we focus on algorithms for efficiently performing arithmetic operations such as addition, multiplication and division, and their connections to topics such as modular arithmetic, greatest common divisors, the Fast Fourier Transform (FFT), and the computation of elementary and special functions. The algorithms that we present are mainly intended for arbitrary-precision arithmetic. They are not limited by the computer word size, only by the memory and time available for the computation. We consider both integer and real (floating-point) computations. The book is divided into four main chapters, plus an appendix. Our aim is to present the latest developments in a concise manner. At the same time, we provide a self-contained introduction for the reader who is not an expert in the field, and exercises at the end of each chapter. Chapter titles are: 1, Integer Arithmetic; 2, Modular Arithmetic and the FFT; 3, Floating-Point Arithmetic; 4, Elementary and Special Function Evaluation; 5 (Appendix), Implementations and Pointers. The book also contains a bibliography of 236 entries, index, summary of notation, and summary of complexities.

cs.DS↗

arXiv2010

Unrestricted algorithms for elementary and special functions

We describe some "unrestricted" algorithms which are useful for the computation of elementary and special functions when the precision required is not known in advance. Several general classes of algorithms are identified and illustrated by examples. The topics include: power series methods, use of halving identities, asymptotic expansions, continued fractions, recurrence relations, Newton's method, numerical contour integration, and the arithmetic-geometric mean. Most of the algorithms discussed are implemented in the MP package (arXiv:1004.3173).

math.NA↗

arXiv2010

Some linear-time algorithms for systolic arrays

We survey some results on linear-time algorithms for systolic arrays. In particular, we show how the greatest common divisor (GCD) of two polynomials of degree n over a finite field can be computed in time O(n) on a linear systolic array of O(n) cells; similarly for the GCD of two n-bit binary numbers. We show how n * n Toeplitz systems of linear equations can be solved in time O(n) on a linear array of O(n) cells, each of which has constant memory size (independent of n). Finally, we outline how a two-dimensional square array of O(n)* O(n) cells can be used to solve (to working accuracy) the eigenvalue problem for a symmetric real n* n matrix in time O(nS(n)). Here S(n) is a slowly growing function of n; for practical purposes S(n) can be regarded as a constant. In addition to their theoretical interest, these results have potential applications in the areas of error-correcting codes, symbolic and algebraic computations, signal processing and image processing.

cs.DS↗

arXiv2010

Fast normal random number generators on vector processors

We consider pseudo-random number generators suitable for vector processors. In particular, we describe vectorised implementations of the Box-Muller and Polar methods, and show that they give good performance on the Fujitsu VP2200. We also consider some other popular methods, e.g. the Ratio method of Kinderman and Monahan (1977) (as improved by Leva (1992)), and the method of Von Neumann and Forsythe, and show why they are unlikely to be competitive with the Polar method on vector processors.

cs.DS↗

arXiv2010

Uses of randomness in computation

Random number generators are widely used in practical algorithms. Examples include simulation, number theory (primality testing and integer factorization), fault tolerance, routing, cryptography, optimization by simulated annealing, and perfect hashing. Complexity theory usually considers the worst-case behaviour of deterministic algorithms, but it can also consider average-case behaviour if it is assumed that the input data is drawn randomly from a given distribution. Rabin popularised the idea of "probabilistic" algorithms, where randomness is incorporated into the algorithm instead of being assumed in the input data. Yao showed that there is a close connection between the complexity of probabilistic algorithms and the average-case complexity of deterministic algorithms. We give examples of the uses of randomness in computation, discuss the contributions of Rabin, Yao and others, and mention some open questions. This is the text of an invited talk presented at "Theory Day", University of NSW, Sydney, 22 April 1994.

cs.DS↗

arXiv2010

Factorizations of Cunningham numbers with bases 13 to 99

This Report updates the tables of factorizations of a^n +- 1 for 13 < a < 100, previously published as CWI Report NM-R9212 (June 1992) and updated in CWI Report NM-R9419 (Update 1, September 1994) and CWI Report NM-R9609 (Update 2, March 1996). A total of 951 new entries in the tables are given here. The factorizations are now complete for n < 76, and there are no composite cofactors smaller than 10^102.

math.NT↗

arXiv2010

MP users guide

MP is a package of ANSI Standard Fortran (ANS X3.9-1966) subroutines for performing multiple-precision floating-point arithmetic and evaluating elementary and special functions. The subroutines are machine independent and the precision is arbitrary, subject to storage limitations. The User's Guide describes the routines and their calling sequences, example and test programs, use of the Augment precompiler, and gives installation instructions for the package.

cs.MS↗

arXiv2010

Some integer factorization algorithms using elliptic curves

Lenstra's integer factorization algorithm is asymptotically one of the fastest known algorithms, and is ideally suited for parallel computation. We suggest a way in which the algorithm can be speeded up by the addition of a second phase. Under some plausible assumptions, the speedup is of order log(p), where p is the factor which is found. In practice the speedup is significant. We mention some refinements which give greater speedup, an alternative way of implementing a second phase, and the connection with Pollard's "p-1" factorization algorithm.

math.NT↗

arXiv2010

On the precision attainable with various floating-point number systems

For scientific computations on a digital computer the set of real number is usually approximated by a finite set F of "floating-point" numbers. We compare the numerical accuracy possible with difference choices of F having approximately the same range and requiring the same word length. In particular, we compare different choices of base (or radix) in the usual floating-point systems. The emphasis is on the choice of F, not on the details of the number representation or the arithmetic, but both rounded and truncated arithmetic are considered. Theoretical results are given, and some simulations of typical floating-point computations (forming sums, solving systems of linear equations, finding eigenvalues) are described. If the leading fraction bit of a normalized base 2 number is not stored explicitly (saving a bit), and the criterion is to minimize the mean square roundoff error, then base 2 is best. If unnormalized numbers are allowed, so the first bit must be stored explicitly, then base 4 (or sometimes base 8) is the best of the usual systems.

math.NA↗

arXiv2010

A fast vectorised implementation of Wallace's normal random number generator

Wallace has proposed a new class of pseudo-random generators for normal variates. These generators do not require a stream of uniform pseudo-random numbers, except for initialisation. The inner loops are essentially matrix-vector multiplications and are very suitable for implementation on vector processors or vector/parallel processors such as the Fujitsu VPP300. In this report we outline Wallace's idea, consider some variations on it, and describe a vectorised implementation RANN4 which is more than three times faster than its best competitors (the Polar and Box-Muller methods) on the Fujitsu VP2200 and VPP300.

cs.DS↗

arXiv2010

Some long-period random number generators using shifts and xors

Marsaglia recently introduced a class of xorshift random number generators (RNGs) with periods 2n-1 for n = 32, 64, etc. Here we give a generalisation of Marsaglia's xorshift generators in order to obtain fast and high-quality RNGs with extremely long periods. RNGs based on primitive trinomials may be unsatisfactory because a trinomial has very small weight. In contrast, our generators can be chosen so that their minimal polynomials have large weight (number of nonzero terms). A computer search using Magma has found good generators for n a power of two up to 4096. These have been implemented in a free software package xorgens.

cs.DS↗