| 摘要 |
PURPOSE:To speedily calculate a floating point by obtaining an approximate value from an approximate value table, and calculating a function value from the approximate value according to an approximate method. CONSTITUTION:The approximate value table is stored in a data RAM. A shifter obtains the number of the top address of the approximate value table up to the approximate value concerned and the approximate value itself. According to the approximate method the function value is obtained from the approximate value. Prescribed function approximate value tables 21, 22 and 23 are stored in the data RAM. When a function F(x) is given, the approximate value table corresponding to the function F(x) is selected. On the other hand, the shifter 5 obtains a number x1 showing the address of the corresponding approximate value from a given variable. Thus the approximate value alpha is obtained, and the function value (y) is calculated based on the approximate value and according to the approximate method.
|
