The MPFUN90 software permits a new or existing Fortran-90 program to utilize multiple-precision arithmetic, or in other words a level of numeric precision that ranges from 50 digits to hundreds or thousands of digits if required. Such high precision Is required by a rapidly expanding body of scientific computations in physics and mathematics, for which the conventional 64-bit IEEE computer arithmetic (about 16 decimal digit accuracy) is not sufficient. The package includes an extensive set of low-level routines to perform high-precision arithmetic, including routines to calculate various algebraic and transcendental functions, such as square roots, sin, cos, exp, log and others. In addition, the package includes high-level translation facilities, so that Fortran programs can utilize these facilities by making only a few changes to conventional Fortran programs. In most cases, the only changes that are required are to change the type statements of variables that one wishes to be treated as multiple precision, plus a few other minor changes. Sample application programs are included for performing polynomial root calculations, high-precision evaluations of numerical integrals, and several versions of the PSLQ integer relation detection algorithm. The MPFUN9O package is based an earlier package written by the author (David H Bailey) when he was at NASA Ames Research Center. It has recently been revamped and updated to be equivalent, from a users perspective, to the Arbitrary Precision (ARPREC) package, which has been produced more recently at Lawrence Berkeley National Laboratory. The chief difference is that the MPFUN9O package is written exclusively in Fortran-90, thus avoiding difficulties that some users experience with ARPREC, which is coded both in C++ and Fortran.
To order this software or receive further information, please fill out the following request: Request Software
@misc{osti_1230773,
title = {A Fortran-90 Based Multiprecision System, Version 00},
author = {Bailey, David H.},
abstractNote = {The MPFUN90 software permits a new or existing Fortran-90 program to utilize multiple-precision arithmetic, or in other words a level of numeric precision that ranges from 50 digits to hundreds or thousands of digits if required. Such high precision Is required by a rapidly expanding body of scientific computations in physics and mathematics, for which the conventional 64-bit IEEE computer arithmetic (about 16 decimal digit accuracy) is not sufficient. The package includes an extensive set of low-level routines to perform high-precision arithmetic, including routines to calculate various algebraic and transcendental functions, such as square roots, sin, cos, exp, log and others. In addition, the package includes high-level translation facilities, so that Fortran programs can utilize these facilities by making only a few changes to conventional Fortran programs. In most cases, the only changes that are required are to change the type statements of variables that one wishes to be treated as multiple precision, plus a few other minor changes. Sample application programs are included for performing polynomial root calculations, high-precision evaluations of numerical integrals, and several versions of the PSLQ integer relation detection algorithm. The MPFUN9O package is based an earlier package written by the author (David H Bailey) when he was at NASA Ames Research Center. It has recently been revamped and updated to be equivalent, from a users perspective, to the Arbitrary Precision (ARPREC) package, which has been produced more recently at Lawrence Berkeley National Laboratory. The chief difference is that the MPFUN9O package is written exclusively in Fortran-90, thus avoiding difficulties that some users experience with ARPREC, which is coded both in C++ and Fortran.},
doi = {},
url = {https://www.osti.gov/biblio/1230773},
year = {Wed Dec 01 00:00:00 EST 2004},
month = {Wed Dec 01 00:00:00 EST 2004},
note =
}