METHOD AND DEVICE FOR HIGH SPEED CALCULATION OF NATURAL LOGARITHM log (X)

摘要

PROBLEM TO BE SOLVED: To reduce complexity of proximity to be used for calculating natural logarithm as achieving numerical accuracy coincident with a floating point accuracy of IEEE by a method for calculating natural logarithmic function. SOLUTION: This method includes a step to divide a mantissa area between 1 and 2 into N small areas with even intervals, a step to preliminarily calculate each center point ax (i=0 to N-1) of the N small areas with even intervals, a step to select a value of N as sufficiently large so that log (m) is calculated within a preliminarily selected accuracy range for an optional m in the small area by a primary polynomial expression of m by defining m as a mantissa for binary floating-point expression for a certain numeral in each small area and a step to calculate a value of log (X) for the binary floating-point expression of a specified numeral x stored in a memory of an operational device by using the primary polynomial expression of m.