Elsevier Science Home
Computer Physics Communications Program Library
Full text online from Science Direct
Programs in Physics & Physical Chemistry
CPC Home

[Licence| Download | New Version Template] abrj_v1_0.gz(16 Kbytes)
Manuscript Title: MP: a multiple precision package.
Authors: O. Portilho
Program title: MP
Catalogue identifier: ABRJ_v1_0
Distribution format: gz
Journal reference: Comput. Phys. Commun. 59(1990)345
Programming language: Fortran.
Computer: VAX 8350/MICROVAX.
Operating system: ULTRIX/VMS.
RAM: 14K words
Word size: 32
Keywords: General purpose, Utility, Precision, Multiple precision Calculations.
Classification: 4.14.

Nature of problem:
In many situations a higher precision than the one furnished by the machine is necessary to get confinding results, like in variational calculations in Nuclear Physics, when large harmonic oscillator bases are used.

Solution method:
Mantissas of numbers are stored in REAL*16 integer constants, each one containing 30 digits, and exponents are stored in INTERGER*4 constants. Using this multiple precision representation of numbers, FORTRAN subroutines were written to perform basic arithmetic operations and to supply commonly required mathematical functions in such a way to provide the interested user a tool to make high precision calculations. The algorithms for each subroutine are depicted in the long write-up.

Running time:
The running time depends on the precision required in the calculations. For the test run, in which 60 digits were employed, a VAX 8350 (ULTRIX) took 46.4 seconds of CPU time and a Microvax (VMS) took 13.4 minutes.