| 摘要 |
A variable floating-point data microprocessor. A floating-point number is composed of a semiotic domain, an exponent domain and a mantissa domain. When a program uses custom floating-point numbers, a custom floating-point configuration register 0, a custom floating-point configuration register 1, a custom floating-point configuration register 2 and a custom floating-point configuration register 3 need to be configured; the order of composition of the custom floating-point numbers is the order of composition of the custom floating-point configuration registers, and the order of composition of the custom floating-point numbers is as follows: the custom floating-point configuration register 3, the custom floating-point configuration register 2, the custom floating-point configuration register 1 and the custom floating point configuration register 0; the sum of the length of the mantissa domain, the length of the exponent domain and the length of the semiotic domain is equal to the length of floating-point data, otherwise, anomalies occur. During a calculation process, the microprocessor can vary the order and lengths of a sign bit, an index bit and a mantissa bit in a floating-point number, and also the width of the floating-point number. Good portability and higher safety are achieved, and floating-point data with higher precisions can be expressed. |
