Floating point arithmetic unit

TL;DR: In this paper, a floating point arithmetic unit which includes a cancelation discrimination is provided, where the possibility of cancelation is recognized by using a BCD(binary coded decimal) operation.

Abstract: PROBLEM TO BE SOLVED: To provide a floating point arithmetic unit which includes a cancelation discrimination means to use a BCD(binary coded decimal) operation if the possibility of cancelation is recognized and therefore obtaining an accurate result including no cancelation error to prevent much deterioration of the processing speed SOLUTION: The arithmetic expression inputted via an arithmetic expression input means 1 is analyzed by an arithmetic expression analysis means 2 in regard to the arithmetic type and the numerical data Then an arithmetic control means 3 analyzes a series of arithmetic expressions of arithmetic operations into combinations of binomial operations When the binomial operations are carried out, a cancelation discrimination means 4 discriminates whether the cancelation occurs or not If the possibility of cancelation is recognized, a data conversion means 6 converts the numerical data into a BCD form and the data of the BCD form are calculated by a BCD calculation means 52 If the possibility of cancelation is not recognized, the numerical data are calculated by a floating point calculation means 51 The obtained arithmetic result is reflected on the original arithmetic expression and the binomial operations are repeated until the arithmetic expression is over