PROCESSOR AND METHOD FOR NUMERICAL DATA ARITHMETIC PROCESSING, VERY SMALL-SIZED ARITHMETIC PROCESSOR, COMPILER DEVICE, AND STORAGE MEDIUM STORED WITH COMPILER

摘要

PROBLEM TO BE SOLVED: To make very small the number of digits of a shift of a mantissa part for equalizing a value to an exponential part by providing a register showing a cardinal part and a register showing the exponential part for floating-point arithmetic processing wherein real number numerical data is represented as a specific expression. ;SOLUTION: The arithmetic processor for numerical data represented as an exponential part and a mantissa part is equipped with the register B showing the cardinal part B and the register E showing the exponetial part E for the floating-point arithmetic processing wherein the real number numeric data R is represented as an expression including M as the mantissa part and X as the base. Then the real number numerical data R is converted as shown by the expression and the floating-point arithmetic process is performed. Further, the very small-sized arithmetic processor which performs arithmetic processing of the numerical data R represented as the exponential part E and mantissa part M is equipped with a register B, a register E, and a control storage memory storing microcodes for controlling the registers B and E for the floating-point arithmetic processing represented by the expression.;COPYRIGHT: (C)1999,JPO