Floating point (IEEE-754)

Floating-point numbers are also called "real numbers", and a floating-point number consists of three parts

1. Sign Bit (S)
2. Order Code
3. Valid numbers

S: Order Code: Valid numbers

Floating-point numbers are represented by the science two-level system.

Three types of floating-point numbers:

1. Short floating point (32bit): S (BIT31): Order Code (BIT30~BIT23): Valid number (BIT22~BIT0)
2. Long floating point (64bit): S (BIT63): Order Code (BIT62~BIT52): Valid number (BIT51~BIT0)
3. Temporary character point (80bit): S (bit79): Order Code (BIT78~BIT64): Valid number (BIT63~BIT0)

Floating-point conversion

1. Convert 10 binary to binary number

2. Normalize the binary number

3. Calculate the Order Code

4. Store this number in floating-point format

For example: floating point 100.2510, to a single-precision floating point number

1.100.25 = 1100100.01

2.1100100.01 = 1.10010001 x 2^6

3.110 + 0111111 = 10000101

4. Sign bit = 0

Order Code = 10000101

Valid number = 1001 0001 0000 0000 0000 000

Floating point (IEEE-754)