Apparatus and method for converting floating point number into integer in floating point unit

Abstract

The present invention relates to an apparatus and a method for converting the binary floating point number into a integer represented as binary numbers of fixed length, the method for converting the binary floating point number into a integer comprising: calculating an exponent difference between the floating point number as a first operand and a constant which has a fraction of `0` value and an exponent of `data sizes of the integer-1` value as a second operand, shifting the fraction of the operand having the smaller exponent as many as the exponent difference, and outputting a bigger exponent; adding the fraction of the operand having the bigger exponent and a selected fraction, wherein the a selected fraction is in case of a negative value, the shifted fraction which is complemented to one, in case of a positive value, the shifted fraction itself in response to a sign of the first operand; and outputting the bigger exponent as a final result of the exponent, outputting and rounding the added fraction in response to a control signal as the final result of the fraction.

Dependencies are inferred from the claim text.