If (x> 0.000001 & X

Float accuracy in C
Float storage format:
S E M
One-digit symbol-bit 8-digit index 23-digit ending number

Convert to a value that is: V = (-1) ^ s * 1.m * 2 ^ (E-127)

If 16.5 is converted to binary 00010000.1 => 1.00001*2 ^ 4,
The memory representation is:
Symbol bit index 4 + 127 = 131 tails
0 10000011 00001 000000000000000000

In the conversion process, it is obvious that the higher the integer part of the float number, the lower the precision of the decimal part.
For the float number, the valid number is about 7 digits (2 ^ 23 is about equal to 10 ^ 7). Therefore, the more digits the integer part occupies
The lower the precision, when the integer part exceeds 9999999, the decimal part is completely inaccurate.