| 摘要 |
PURPOSE:To extend the range of numeric data by equipping a data processor which processes floating-point data with a logarithm transform processing part converting floating-point data into logarithms and by performing a process by the logarithm floating-point data of the floaing-poing data. CONSTITUTION:When two kinds of floating-point data X and Y and two kinds of logarithmic-value floating data (log x) and (log y) are given, their operand data are set to registers 1 and 2 respectively and converted by logarithm transform processing parts 2 and 3 inverse transform processing parts 5 and 6 into pieces of data X and Y, and (log x) and (log y) which are inputted to arithmetic circuit part 8. This arithmetic circuit part 8 is provided with floating-point data arithemetic circuit 9, logarithmic-value floating-point data arithmetic circuit 10 and overflow detection part 11 detecting an overflow, so that gates 19 and 20 will be controlled when an overflow or underflow occurs. |
