METHOD FOR PROCESSING EXPONENTS ON MULTIPLICATION/DIVISION OPERATION OF FLOATING POINT OPERATION

摘要

PURPOSE: A method for processing exponents on a multiplication/division operation of a floating point operation is provided to reduce an unnecessary logic and increase speed on entire operation by making hardware perform a bias correction problem generated when the exponent is respectively added or subtracted on the multiplication/division. CONSTITUTION: The biased exponents are generated by respectively adding the bias fixed to the exponents. An adder adds the biased exponents. A fixed bias value is subtracted from an addition result of the biased exponents. In case that the fixed bias value is 1, bias correction is preformed by making an initial carry input terminal of the adder performing the addition receive a signal informing an operator of the addition.