Design and Implementation of a High Precision Arithmetic with Rigorous Error Bounds

Abstract

In this paper we present the design of a rigorous, high precision floating point arithmetic. The algorithms presented are implemented in FORTRAN, and are used in a special version of the COSY INFINITY rigorous computation package.

The three design objectives for these high precision intervals are high speed, particularly for the elementary operations, absolutely rigorous treatment of round- off errors, and scalability to arbitrary precision. The main focus of these algo- rithms lies on relatively low precision of up to 100 significant digits.

Unlike many available high precision packages, we do not rely on large integer arithmetic to emulate floating point numbers with arbitrary mantissa length. Instead, we utilize unevaluated series of double precision floating point numbers. Combined with algorithms for exact operations on floating point numbers, this allows us to implement operations very efficiently.

A. Wittig, M. Berz, MSUHEP-081126, Michigan State University (2008)

Download

Download Adobe PDF version (431088 Bytes).

Go Back to the reprint server.
Go Back to the home page.

This page is maintained by Kyoko Makino. Please contact her if there are any problems with it.