| 摘要 |
PURPOSE:To increase an arithmetic speed through a simple circuit by performing conversion from a binary number to an absolute value or reverse conversion through a (A-1)' arithmetic circuit when conversion from a fixed point number to a floating point number or reverse conversion is performed. CONSTITUTION:When a positive fixed point number is converted into a floating point number, the positive fixed point number X is inputted to an input side firstly. Consequently, the contents of an input sign register 1 is set to 0, and the contents of an input exponent part register are all set to 0, and X is stored in an input mantissa part register 3 to a width of the low-order 32 bits. A floating/fixed point indicating circuit 6, on the other hand, outputs 0, and then the 1st selecting circuit 7 selects the output side of the input mantissa part register 3 and supplies it to an arithmetic circuit 8 and one input side of the 2nd selecting circuit 9. When the number is positive, 0 as the contents of the input sign register 1 is supplied to and stored in an output sign register 13, 0 is supplied to the 2nd selecting circuit 9, and consequently the output of the 1st selecting circuit 7 is selected and supplied to the 3rd selecting circuit 11 and the 4th selecting circuit 12 respectively. |
