Floating point and integer number conversions in a floating point adder

Apparatus for converting from a floating point number which includes a mantissa field, an exponent field, and a sign field to an integer of n-bit size including a sign bit, in accordance with a control word specifying a particular rounding mode, comprising:

• floating point number storage means for storing said floating point number;
• control word storage means for storing said control word;
• first means for shifting said mantissa field right to thereby denormalize said mantissa resulting in a shifted mantissa;
• second means connected to said first means for subtracting zero from said shifted mantissa to produce a first result mantissa field that is a two's complement number, upon the condition that said sign field indicates that said floating point number is a negative number;
• said second means including third means for adding zero to said shifted mantissa to produce a second result mantissa field, upon the condition that said sign field indicates that said floating point number is a positive number;
• fourth means connected to said second means for detecting overflow of said result mantissa field with respect to said integer of n-bit size;
• fifth means connected to said floating point number storage means and to said second means for setting said sign bit of said integer to reflect the fact that said sign field of said floating point number is a negative or positive number, and,
• sixth means connected to said second means and to said control word storage means for rounding said integer in accordance with a rounding mode specified in said control word provided that said sign field indicates that said floating point number is a positive number and for rounding said integer in accordance with a logic table provided that said sign field indicates that said floating point number is a negative number.