Resolving IEEE floating-point error using precision-based rounding algorithm

Abstract

Thispaperpresents a precision-based rounding approach toresolve IEEEfloating- pointerror(1). Binary floating-point systems canrepresent onlyafinite numberoffloating- point values inexact form. Allother values must beapproximated bytoclosest representable value. Theproposed method notonly provides a precision-based rounding, butalsorepresents floating-point values inexact formapplicable to anyarithmetic operation. Theerror estimation accuracy isevaluated across several different applications. However, theaccuracy slightly decreases withthedepth ofprecision. Thegoal istopresent asolution forIEEEfloating-point error forcommonly usedarithmetic operations usedinbusiness applications. Thissimple cprograms explains theabove mentioned problem #includ intmain()