Computing predecessor and successor in rounding to nearest (Q2391021)

The authors present a substantial improvement for interval operations giving an estimation of the predecessor and successor of a floating-point number using only a floating-point operations in rounding to nearest. The rounding to nearest is defined according to the IEEE 754 standard and also to the IEEE 754-2008 revised standard. The essential contribution of this paper is a simple and efficient algorithm for any radix to compute the bounds of an interval containing the floating arithmetical operations.