Algorithm 707: CONHYP: a numerical evaluator of the confluent hypergeometric function for complex arguments of large magnitudes

Authors:
Mark Nardin;W. F. Perger;Atul Bhalla
Affiliations:
University of Michigan, Ann Arbor;Michigan Technological University;Michigan Technological University
Venue:
ACM Transactions on Mathematical Software (TOMS)
Year:
1992

Citing 2
Cited 2

Numerical evaluation of the confluent hypergeometric function for complex arguments of large magnitudes

Journal of Computational and Applied Mathematics
Handbook of Mathematical Functions, With Formulas, Graphs, and Mathematical Tables,

Handbook of Mathematical Functions, With Formulas, Graphs, and Mathematical Tables,

Moment and Hypergeometric Filters for High Precision Computation ofFocus, Stereo and Optical Flow

International Journal of Computer Vision
Hypergeometric Filters for Optical Flow and Affine Matching

International Journal of Computer Vision

Quantified Score

Hi-index	0.00

Visualization

Abstract

A numerical evaluator for the confluent hypergeometric function for complex arguments with large magnitudes using a direct summation of Kummer's series is presented. Extended precision subroutines using large arrays to accumulate a single numerator and denominator are ultimately used in a single division to arrive at the final result. The accuracy has been verified through a variety of tests and they show the evaluator to be consistently accurate to 13 significant figures, and on rare occasion accurate to only 9 for magnitudes of the arguments ranging into the thousands in any quadrant in the complex plane. Because the evaluator automatically determines the number of significant figures of machine precision, and because it is written in FORTRAN 77, tests on various computers have shown the evaluator to provide consistently accurate results, making the evaluator very portable. The principal drawback is that, for certain arguments, the evaluator is slow; however, the evaluator remains valuable as a benchmark even in such cases.